12345...102030...


AI File Extension – What is a .ai file and how do I open it?

An AI file is a drawing created with Adobe Illustrator, a vector graphics editing program. It is composed of paths connected by points, rather than bitmap image data. AI files are commonly used for logos and print media.

AI file open in Adobe Illustrator CC 2017

Since Illustrator image files are saved in a vector format, they can be enlarged without losing any image quality. Some third-party programs can open AI files, but they may rasterize the image, meaning the vector data will be converted to a bitmap format.

To open an Illustrator document in Photoshop, the file must first have PDF Content saved within the file. If it does not contain the PDF Content, then the graphic cannot be opened and will display a default message, stating, “This is an Adobe Illustrator file that was saved without PDF Content. To place or open this file in other applications, it should be re-saved from Adobe Illustrator with the “Create PDF Compatible File” option turned on.”

See the rest here:

AI File Extension – What is a .ai file and how do I open it?

Posted in Ai

Ideas about Ai

Subscribe to receive email notificationswhenever new talks are published.

Please enter an email address.

Please enter a valid email address.

Did you mean?

Please checkDailyorWeeklyand try again.

Please check your details and try again.

Please check your details and try again.

Sorry, we’re currently having troubleprocessing new newsletter signups.Please try again later.

View original post here:

Ideas about Ai

Posted in Ai

AI File – What is it and how do I open it?

Did your computer fail to open an AI file? We explain what AI files are and recommend software that we know can open or convert your AI files.

AI is the acronym for Adobe Illustrator. Files that have the .ai extension are drawing files that the Adobe Illustrator application has created.

The Adobe Illustrator application was developed by Adobe Systems. The files created by this application are composed of paths that are connected by points and are saved in vector format. The technology used to create these files allows the user to re-size the AI image without losing any of the image’s quality.

Some third-party programs allow users to “rastersize” the images created in Adobe Illustrator, which allows them to convert the AI file into bitmap format. While this may make the file size smaller and easier to open across multiple applications, some of the file quality may be lost in the process.

Go here to see the original:

AI File – What is it and how do I open it?

Posted in Ai

Artificial intelligence – Wikipedia

Artificial intelligence (AI, also machine intelligence, MI) is intelligence demonstrated by machines, in contrast to the natural intelligence (NI) displayed by humans and other animals. In computer science AI research is defined as the study of “intelligent agents”: any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.[1] Colloquially, the term “artificial intelligence” is applied when a machine mimics “cognitive” functions that humans associate with other human minds, such as “learning” and “problem solving”.

The scope of AI is disputed: as machines become increasingly capable, tasks considered as requiring “intelligence” are often removed from the definition, a phenomenon known as the AI effect, leading to the quip, “AI is whatever hasn’t been done yet.”[3] For instance, optical character recognition is frequently excluded from “artificial intelligence”, having become a routine technology.[4] Capabilities generally classified as AI as of 2017[update] include successfully understanding human speech, competing at the highest level in strategic game systems (such as chess and Go[6]), autonomous cars, intelligent routing in content delivery network and military simulations.

Artificial intelligence was founded as an academic discipline in 1956, and in the years since has experienced several waves of optimism,[7][8] followed by disappointment and the loss of funding (known as an “AI winter”),[9][10] followed by new approaches, success and renewed funding.[8][11] For most of its history, AI research has been divided into subfields that often fail to communicate with each other.[12] These sub-fields are based on technical considerations, such as particular goals (e.g. “robotics” or “machine learning”),[13] the use of particular tools (“logic” or “neural networks”), or deep philosophical differences.[14][15][16] Subfields have also been based on social factors (particular institutions or the work of particular researchers).[12]

The traditional problems (or goals) of AI research include reasoning, knowledge representation, planning, learning, natural language processing, perception and the ability to move and manipulate objects.[13] General intelligence is among the field’s long-term goals.[17] Approaches include statistical methods, computational intelligence, and traditional symbolic AI. Many tools are used in AI, including versions of search and mathematical optimization, neural networks and methods based on statistics, probability and economics. The AI field draws upon computer science, mathematics, psychology, linguistics, philosophy and many others.

The field was founded on the claim that human intelligence “can be so precisely described that a machine can be made to simulate it”.[18] This raises philosophical arguments about the nature of the mind and the ethics of creating artificial beings endowed with human-like intelligence, issues which have been explored by myth, fiction and philosophy since antiquity.[19] Some people also consider AI to be a danger to humanity if it progresses unabatedly.[20] Others believe that AI, unlike previous technological revolutions, will create a risk of mass unemployment.[21]

In the twenty-first century, AI techniques have experienced a resurgence following concurrent advances in computer power, large amounts of data, and theoretical understanding; and AI techniques have become an essential part of the technology industry, helping to solve many challenging problems in computer science.[22][11]

Thought-capable artificial beings appeared as storytelling devices in antiquity,[23] and have been common in fiction, as in Mary Shelley’s Frankenstein or Karel apek’s R.U.R. (Rossum’s Universal Robots).[24] These characters and their fates raised many of the same issues now discussed in the ethics of artificial intelligence.[19]

The study of mechanical or “formal” reasoning began with philosophers and mathematicians in antiquity. The study of mathematical logic led directly to Alan Turing’s theory of computation, which suggested that a machine, by shuffling symbols as simple as “0” and “1”, could simulate any conceivable act of mathematical deduction. This insight, that digital computers can simulate any process of formal reasoning, is known as the ChurchTuring thesis.[25] Along with concurrent discoveries in neurobiology, information theory and cybernetics, this led researchers to consider the possibility of building an electronic brain. Turing proposed that “if a human could not distinguish between responses from a machine and a human, the machine could be considered intelligent”.[26] The first work that is now generally recognized as AI was McCullouch and Pitts’ 1943 formal design for Turing-complete “artificial neurons”.

The field of AI research was born at a workshop at Dartmouth College in 1956.[28] Attendees Allen Newell (CMU), Herbert Simon (CMU), John McCarthy (MIT), Marvin Minsky (MIT) and Arthur Samuel (IBM) became the founders and leaders of AI research.[29] They and their students produced programs that the press described as “astonishing”: computers were learning checkers strategies (c. 1954)[31] (and by 1959 were reportedly playing better than the average human),[32] solving word problems in algebra, proving logical theorems (Logic Theorist, first run c. 1956) and speaking English.[33] By the middle of the 1960s, research in the U.S. was heavily funded by the Department of Defense[34] and laboratories had been established around the world.[35] AI’s founders were optimistic about the future: Herbert Simon predicted, “machines will be capable, within twenty years, of doing any work a man can do”. Marvin Minsky agreed, writing, “within a generation… the problem of creating ‘artificial intelligence’ will substantially be solved”.[7]

They failed to recognize the difficulty of some of the remaining tasks. Progress slowed and in 1974, in response to the criticism of Sir James Lighthill and ongoing pressure from the US Congress to fund more productive projects, both the U.S. and British governments cut off exploratory research in AI. The next few years would later be called an “AI winter”,[9] a period when obtaining funding for AI projects was difficult.

In the early 1980s, AI research was revived by the commercial success of expert systems,[37] a form of AI program that simulated the knowledge and analytical skills of human experts. By 1985 the market for AI had reached over a billion dollars. At the same time, Japan’s fifth generation computer project inspired the U.S and British governments to restore funding for academic research.[8] However, beginning with the collapse of the Lisp Machine market in 1987, AI once again fell into disrepute, and a second, longer-lasting hiatus began.[10]

In the late 1990s and early 21st century, AI began to be used for logistics, data mining, medical diagnosis and other areas.[22] The success was due to increasing computational power (see Moore’s law), greater emphasis on solving specific problems, new ties between AI and other fields (such as statistics, economics and mathematics), and a commitment by researchers to mathematical methods and scientific standards.[38] Deep Blue became the first computer chess-playing system to beat a reigning world chess champion, Garry Kasparov on 11 May 1997.

In 2011, a Jeopardy! quiz show exhibition match, IBM’s question answering system, Watson, defeated the two greatest Jeopardy champions, Brad Rutter and Ken Jennings, by a significant margin. Faster computers, algorithmic improvements, and access to large amounts of data enabled advances in machine learning and perception; data-hungry deep learning methods started to dominate accuracy benchmarks around 2012.[41] The Kinect, which provides a 3D bodymotion interface for the Xbox 360 and the Xbox One use algorithms that emerged from lengthy AI research[42] as do intelligent personal assistants in smartphones.[43] In March 2016, AlphaGo won 4 out of 5 games of Go in a match with Go champion Lee Sedol, becoming the first computer Go-playing system to beat a professional Go player without handicaps.[6][44] In the 2017 Future of Go Summit, AlphaGo won a three-game match with Ke Jie,[45] who at the time continuously held the world No. 1 ranking for two years.[46][47] This marked the completion of a significant milestone in the development of Artificial Intelligence as Go is an extremely complex game, more so than Chess.

According to Bloomberg’s Jack Clark, 2015 was a landmark year for artificial intelligence, with the number of software projects that use AI within Google increased from a “sporadic usage” in 2012 to more than 2,700 projects. Clark also presents factual data indicating that error rates in image processing tasks have fallen significantly since 2011.[48] He attributes this to an increase in affordable neural networks, due to a rise in cloud computing infrastructure and to an increase in research tools and datasets.[11] Other cited examples include Microsoft’s development of a Skype system that can automatically translate from one language to another and Facebook’s system that can describe images to blind people.[48]

A typical AI perceives its environment and takes actions that maximize its chance of successfully achieving its goals.[1] An AI’s intended goal function can be simple (“1 if the AI wins a game of Go, 0 otherwise”) or complex (“Do actions mathematically similar to the actions that got you rewards in the past”). Goals can be explicitly defined, or can be induced. If the AI is programmed for “reinforcement learning”, goals can be implicitly induced by rewarding some types of behavior and punishing others.[a] Alternatively, an evolutionary system can induce goals by using a “fitness function” to mutate and preferentially replicate high-scoring AI systems; this is similar to how animals evolved to innately desire certain goals such as finding food, or how dogs can be bred via artificial selection to possess desired traits. Some AI systems, such as nearest-neighbor, instead reason by analogy; these systems are not generally given goals, except to the degree that goals are somehow implicit in their training data. Such systems can still be benchmarked if the non-goal system is framed as a system whose “goal” is to successfully accomplish its narrow classification task.[51]

AI often revolves around the use of algorithms. An algorithm is a set of unambiguous instructions that a mechanical computer can execute.[b] A complex algorithm is often built on top of other, simpler, algorithms. A simple example of an algorithm is the following recipe for optimal play at tic-tac-toe:

Many AI algorithms are capable of learning from data; they can enhance themselves by learning new heuristics (strategies, or “rules of thumb”, that have worked well in the past), or can themselves write other algorithms. Some of the “learners” described below, including Bayesian networks, decision trees, and nearest-neighbor, could theoretically, if given infinite data, time, and memory, learn to approximate any function, including whatever combination of mathematical functions would best describe the entire world. These learners could therefore, in theory, derive all possible knowledge, by considering every possible hypothesis and matching it against the data. In practice, it is almost never possible to consider every possibility, because of the phenomenon of “combinatorial explosion”, where the amount of time needed to solve a problem grows exponentially. Much of AI research involves figuring out how to identify and avoid considering broad swaths of possibililities that are unlikely to be fruitful.[53] For example, when viewing a map and looking for the shortest driving route from Denver to New York in the East, one can in most cases skip looking at any path through San Francisco or other areas far to the West; thus, an AI wielding an pathfinding algorithm like A* can avoid the combinatorial explosion that would ensue if every possible route had to be ponderously considered in turn.[55]

The earliest (and easiest to understand) approach to AI was symbolism (such as formal logic): “If an otherwise healthy adult has a fever, then they may have influenza”. A second, more general, approach is Bayesian inference: “If the current patient has a fever, adjust the probability they have influenza in such-and-such way”. The third major approach, extremely popular in routine business AI applications, is analogizers such as SVM and nearest-neighbor: “After examining the records of known past patients whose temperature, symptoms, age, and other factors mostly match the current patient, X% of those patients turned out to have influenza”. A fourth approach is harder to intuitively understand, but is inspired by how the brain’s machinery works: the neural network approach uses artificial “neurons” that can learn by comparing itself to the desired output and altering the strengths of the connections between its internal neurons to “reinforce” connections that seemed to be useful. These four main approaches can overlap with each other and with evolutionary systems; for example, neural nets can learn to make inferences, to generalize, and to make analogies. Some systems implicitly or explicitly use multiple of these approaches, alongside many other AI and non-AI algorithms; the best approach is often different depending on the problem.[57]

Learning algorithms work on the basis that strategies, algorithms, and inferences that worked well in the past are likely to continue working well in the future. These inferences can be obvious, such as “since the sun rose every morning for the last 10,000 days, it will probably rise tomorrow morning as well”. They can be nuanced, such as “X% of families have geographically separate species with color variants, so there is an Y% chance that undiscovered black swans exist”. Learners also work on the basis of “Occam’s razor”: The simplest theory that explains the data is the likeliest. Therefore, to be successful, a learner must be designed such that it prefers simpler theories to complex theories, except in cases where the complex theory is proven substantially better. Settling on a bad, overly complex theory gerrymandered to fit all the past training data is known as overfitting. Many systems attempt to reduce overfitting by rewarding a theory in accordance with how well it fits the data, but penalizing the theory in accordance with how complex the theory is. Besides classic overfitting, learners can also disappoint by “learning the wrong lesson”. A toy example is that an image classifier trained only on pictures of brown horses and black cats might conclude that all brown patches are likely to be horses. A real-world example is that, unlike humans, current image classifiers don’t determine the spatial relationship between components of the picture; instead, they learn abstract patterns of pixels that humans are oblivious to, but that linearly correlate with images of certain types of real objects. Faintly superimposing such a pattern on a legitimate image results in an “adversarial” image that the system misclassifies.[c][60][61][62]

Compared with humans, existing AI lacks several features of human “commonsense reasoning”; most notably, humans have powerful mechanisms for reasoning about “nave physics” such as space, time, and physical interactions. This enables even young children to easily make inferences like “If I roll this pen off a table, it will fall on the floor”. Humans also have a powerful mechanism of “folk psychology” that helps them to interpret natural-language sentences such as “The city councilmen refused the demonstrators a permit because they advocated violence”. (A generic AI has difficulty inferring whether the councilmen or the demonstrators are the ones alleged to be advocating violence.)[65][66][67] This lack of “common knowledge” means that AI often makes different mistakes than humans make, in ways that can seem incomprehensible. For example, existing self-driving cars cannot reason about the location nor the intentions of pedestrians in the exact way that humans do, and instead must use non-human modes of reasoning to avoid accidents.[68][69][70]

The overall research goal of artificial intelligence is to create technology that allows computers and machines to function in an intelligent manner. The general problem of simulating (or creating) intelligence has been broken down into sub-problems. These consist of particular traits or capabilities that researchers expect an intelligent system to display. The traits described below have received the most attention.[13]

Early researchers developed algorithms that imitated step-by-step reasoning that humans use when they solve puzzles or make logical deductions.[71] By the late 1980s and 1990s, AI research had developed methods for dealing with uncertain or incomplete information, employing concepts from probability and economics.[72]

These algorithms proved to be insufficient for solving large reasoning problems, because they experienced a “combinatorial explosion”: they became exponentially slower as the problems grew larger.[53] In fact, even humans rarely use the step-by-step deduction that early AI research was able to model. They solve most of their problems using fast, intuitive judgements.[73]

Knowledge representation[74] and knowledge engineering[75] are central to classical AI research. Some “expert systems” attempt to gather together explicit knowledge possessed by experts in some narrow domain. In addition, some projects attempt to gather the “commonsense knowledge” known to the average person into a database containing extensive knowledge about the world. Among the things a comprehensive commonsense knowledge base would contain are: objects, properties, categories and relations between objects;[76] situations, events, states and time;[77] causes and effects;[78] knowledge about knowledge (what we know about what other people know);[79] and many other, less well researched domains. A representation of “what exists” is an ontology: the set of objects, relations, concepts, and properties formally described so that software agents can interpret them. The semantics of these are captured as description logic concepts, roles, and individuals, and typically implemented as classes, properties, and individuals in the Web Ontology Language.[80] The most general ontologies are called upper ontologies, which attempt to provide a foundation for all other knowledge[81] by acting as mediators between domain ontologies that cover specific knowledge about a particular knowledge domain (field of interest or area of concern). Such formal knowledge representations can be used in content-based indexing and retrieval,[82] scene interpretation,[83] clinical decision support,[84] knowledge discovery (mining “interesting” and actionable inferences from large databases),[85] and other areas.[86]

Among the most difficult problems in knowledge representation are:

Intelligent agents must be able to set goals and achieve them.[93] They need a way to visualize the futurea representation of the state of the world and be able to make predictions about how their actions will change itand be able to make choices that maximize the utility (or “value”) of available choices.[94]

In classical planning problems, the agent can assume that it is the only system acting in the world, allowing the agent to be certain of the consequences of its actions.[95] However, if the agent is not the only actor, then it requires that the agent can reason under uncertainty. This calls for an agent that can not only assess its environment and make predictions, but also evaluate its predictions and adapt based on its assessment.[96]

Multi-agent planning uses the cooperation and competition of many agents to achieve a given goal. Emergent behavior such as this is used by evolutionary algorithms and swarm intelligence.[97]

Machine learning, a fundamental concept of AI research since the field’s inception,[98] is the study of computer algorithms that improve automatically through experience.[99][100]

Unsupervised learning is the ability to find patterns in a stream of input. Supervised learning includes both classification and numerical regression. Classification is used to determine what category something belongs in, after seeing a number of examples of things from several categories. Regression is the attempt to produce a function that describes the relationship between inputs and outputs and predicts how the outputs should change as the inputs change. In reinforcement learning[101] the agent is rewarded for good responses and punished for bad ones. The agent uses this sequence of rewards and punishments to form a strategy for operating in its problem space.

Natural language processing[102] gives machines the ability to read and understand human language. A sufficiently powerful natural language processing system would enable natural language user interfaces and the acquisition of knowledge directly from human-written sources, such as newswire texts. Some straightforward applications of natural language processing include information retrieval, text mining, question answering[103] and machine translation.[104] Many current approaches use word co-occurrence frequencies to construct syntactic representations of text. “Keyword spotting” strategies for search are popular and scalable but dumb; a search query for “dog” might only match documents with the literal word “dog” and miss a document with the word “poodle”. “Lexical affinity” strategies use the occurrence of words such as “accident” to assess the sentiment of a document. Modern statistical NLP approaches can combine all these strategies as well as others, and often achieve acceptable accuracy at the page or paragraph level, but continue to lack the semantic understanding required to classify isolated sentences well. Besides the usual difficulties with encoding semantic commonsense knowledge, existing semantic NLP sometimes scales too poorly to be viable in business applications. Beyond semantic NLP, the ultimate goal of “narrative” NLP is to embody a full understanding of commonsense reasoning.[105]

Machine perception[106] is the ability to use input from sensors (such as cameras, microphones, tactile sensors, sonar and others) to deduce aspects of the world. Computer vision[107] is the ability to analyze visual input. A few selected subproblems are speech recognition,[108] facial recognition and object recognition.[109]

The field of robotics[110] is closely related to AI. Intelligence is required for robots to handle tasks such as object manipulation[111] and navigation, with sub-problems such as localization, mapping, and motion planning. These systems require that an agent is able to: Be spatially cognizant of its surroundings, learn from and build a map of its environment, figure out how to get from one point in space to another, and execute that movement (which often involves compliant motion, a process where movement requires maintaining physical contact with an object).[113]

Within developmental robotics, developmental learning approaches are elaborated upon to allow robots to accumulate repertoires of novel skills through autonomous self-exploration, social interaction with human teachers, and the use of guidance mechanisms (active learning, maturation, motor synergies, etc.).[114][115]

Affective computing is the study and development of systems that can recognize, interpret, process, and simulate human affects. It is an interdisciplinary field spanning computer sciences, psychology, and cognitive science. While the origins of the field may be traced as far back as the early philosophical inquiries into emotion, the more modern branch of computer science originated with Rosalind Picard’s 1995 paper on “affective computing”.[124][125] A motivation for the research is the ability to simulate empathy, where the machine would be able to interpret human emotions and adapts its behavior to give an appropriate response to those emotions.

Emotion and social skills[126] are important to an intelligent agent for two reasons. First, being able to predict the actions of others by understanding their motives and emotional states allow an agent to make better decisions. Concepts such as game theory, decision theory, necessitate that an agent be able to detect and model human emotions. Second, in an effort to facilitate humancomputer interaction, an intelligent machine may want to display emotions (even if it does not experience those emotions itself) to appear more sensitive to the emotional dynamics of human interaction.

Many researchers think that their work will eventually be incorporated into a machine with artificial general intelligence, combining all the skills mentioned above and even exceeding human ability in most or all these areas.[17][127] A few believe that anthropomorphic features like artificial consciousness or an artificial brain may be required for such a project.[128][129]

Many of the problems above may also require general intelligence, if machines are to solve the problems as well as people do. For example, even specific straightforward tasks, like machine translation, require that a machine read and write in both languages (NLP), follow the author’s argument (reason), know what is being talked about (knowledge), and faithfully reproduce the author’s original intent (social intelligence). A problem like machine translation is considered “AI-complete”, because all of these problems need to be solved simultaneously in order to reach human-level machine performance.

There is no established unifying theory or paradigm that guides AI research. Researchers disagree about many issues.[130] A few of the most long standing questions that have remained unanswered are these: should artificial intelligence simulate natural intelligence by studying psychology or neurobiology? Or is human biology as irrelevant to AI research as bird biology is to aeronautical engineering?[14] Can intelligent behavior be described using simple, elegant principles (such as logic or optimization)? Or does it necessarily require solving a large number of completely unrelated problems?[15] Can intelligence be reproduced using high-level symbols, similar to words and ideas? Or does it require “sub-symbolic” processing?[16] John Haugeland, who coined the term GOFAI (Good Old-Fashioned Artificial Intelligence), also proposed that AI should more properly be referred to as synthetic intelligence,[131] a term which has since been adopted by some non-GOFAI researchers.

Stuart Shapiro divides AI research into three approaches, which he calls computational psychology, computational philosophy, and computer science. Computational psychology is used to make computer programs that mimic human behavior.[134] Computational philosophy, is used to develop an adaptive, free-flowing computer mind.[134] Implementing computer science serves the goal of creating computers that can perform tasks that only people could previously accomplish.[134] Together, the humanesque behavior, mind, and actions make up artificial intelligence.

In the 1940s and 1950s, a number of researchers explored the connection between neurobiology, information theory, and cybernetics. Some of them built machines that used electronic networks to exhibit rudimentary intelligence, such as W. Grey Walter’s turtles and the Johns Hopkins Beast. Many of these researchers gathered for meetings of the Teleological Society at Princeton University and the Ratio Club in England.[135] By 1960, this approach was largely abandoned, although elements of it would be revived in the 1980s.

When access to digital computers became possible in the middle 1950s, AI research began to explore the possibility that human intelligence could be reduced to symbol manipulation. The research was centered in three institutions: Carnegie Mellon University, Stanford and MIT, and each one developed its own style of research. John Haugeland named these approaches to AI “good old fashioned AI” or “GOFAI”.[136] During the 1960s, symbolic approaches had achieved great success at simulating high-level thinking in small demonstration programs. Approaches based on cybernetics or neural networks were abandoned or pushed into the background.[137] Researchers in the 1960s and the 1970s were convinced that symbolic approaches would eventually succeed in creating a machine with artificial general intelligence and considered this the goal of their field.

Economist Herbert Simon and Allen Newell studied human problem-solving skills and attempted to formalize them, and their work laid the foundations of the field of artificial intelligence, as well as cognitive science, operations research and management science. Their research team used the results of psychological experiments to develop programs that simulated the techniques that people used to solve problems. This tradition, centered at Carnegie Mellon University would eventually culminate in the development of the Soar architecture in the middle 1980s.[138][139]

Unlike Newell and Simon, John McCarthy felt that machines did not need to simulate human thought, but should instead try to find the essence of abstract reasoning and problem solving, regardless of whether people used the same algorithms.[14] His laboratory at Stanford (SAIL) focused on using formal logic to solve a wide variety of problems, including knowledge representation, planning and learning.[140] Logic was also the focus of the work at the University of Edinburgh and elsewhere in Europe which led to the development of the programming language Prolog and the science of logic programming.[141]

Researchers at MIT (such as Marvin Minsky and Seymour Papert)[142] found that solving difficult problems in vision and natural language processing required ad-hoc solutions they argued that there was no simple and general principle (like logic) that would capture all the aspects of intelligent behavior. Roger Schank described their “anti-logic” approaches as “scruffy” (as opposed to the “neat” paradigms at CMU and Stanford).[15] Commonsense knowledge bases (such as Doug Lenat’s Cyc) are an example of “scruffy” AI, since they must be built by hand, one complicated concept at a time.[143]

When computers with large memories became available around 1970, researchers from all three traditions began to build knowledge into AI applications.[144] This “knowledge revolution” led to the development and deployment of expert systems (introduced by Edward Feigenbaum), the first truly successful form of AI software.[37] The knowledge revolution was also driven by the realization that enormous amounts of knowledge would be required by many simple AI applications.

By the 1980s progress in symbolic AI seemed to stall and many believed that symbolic systems would never be able to imitate all the processes of human cognition, especially perception, robotics, learning and pattern recognition. A number of researchers began to look into “sub-symbolic” approaches to specific AI problems.[16] Sub-symbolic methods manage to approach intelligence without specific representations of knowledge.

This includes embodied, situated, behavior-based, and nouvelle AI. Researchers from the related field of robotics, such as Rodney Brooks, rejected symbolic AI and focused on the basic engineering problems that would allow robots to move and survive.[145] Their work revived the non-symbolic viewpoint of the early cybernetics researchers of the 1950s and reintroduced the use of control theory in AI. This coincided with the development of the embodied mind thesis in the related field of cognitive science: the idea that aspects of the body (such as movement, perception and visualization) are required for higher intelligence.

Interest in neural networks and “connectionism” was revived by David Rumelhart and others in the middle of the 1980s.[146] Neural networks are an example of soft computing — they are solutions to problems which cannot be solved with complete logical certainty, and where an approximate solution is often sufficient. Other soft computing approaches to AI include fuzzy systems, evolutionary computation and many statistical tools. The application of soft computing to AI is studied collectively by the emerging discipline of computational intelligence.[147]

In the 1990s, AI researchers developed sophisticated mathematical tools to solve specific subproblems. These tools are truly scientific, in the sense that their results are both measurable and verifiable, and they have been responsible for many of AI’s recent successes. The shared mathematical language has also permitted a high level of collaboration with more established fields (like mathematics, economics or operations research). Stuart Russell and Peter Norvig describe this movement as nothing less than a “revolution” and “the victory of the neats”.[38] Critics argue that these techniques (with few exceptions) are too focused on particular problems and have failed to address the long-term goal of general intelligence. There is an ongoing debate about the relevance and validity of statistical approaches in AI, exemplified in part by exchanges between Peter Norvig and Noam Chomsky.

In the course of 60 or so years of research, AI has developed a large number of tools to solve the most difficult problems in computer science. A few of the most general of these methods are discussed below.

Many problems in AI can be solved in theory by intelligently searching through many possible solutions:[155] Reasoning can be reduced to performing a search. For example, logical proof can be viewed as searching for a path that leads from premises to conclusions, where each step is the application of an inference rule.[156] Planning algorithms search through trees of goals and subgoals, attempting to find a path to a target goal, a process called means-ends analysis.[157] Robotics algorithms for moving limbs and grasping objects use local searches in configuration space.[111] Many learning algorithms use search algorithms based on optimization.

Simple exhaustive searches[158] are rarely sufficient for most real world problems: the search space (the number of places to search) quickly grows to astronomical numbers. The result is a search that is too slow or never completes. The solution, for many problems, is to use “heuristics” or “rules of thumb” that prioritize choices in favor of those that are more likely to reach a goal, and to do so in a shorter number of steps. In some search methodologies heuristics can also serve to entirely eliminate some choices that are unlikely to lead to a goal (called “pruning the search tree”). Heuristics supply the program with a “best guess” for the path on which the solution lies.[159] Heuristics limit the search for solutions into a smaller sample size.

A very different kind of search came to prominence in the 1990s, based on the mathematical theory of optimization. For many problems, it is possible to begin the search with some form of a guess and then refine the guess incrementally until no more refinements can be made. These algorithms can be visualized as blind hill climbing: we begin the search at a random point on the landscape, and then, by jumps or steps, we keep moving our guess uphill, until we reach the top. Other optimization algorithms are simulated annealing, beam search and random optimization.[160]

Evolutionary computation uses a form of optimization search. For example, they may begin with a population of organisms (the guesses) and then allow them to mutate and recombine, selecting only the fittest to survive each generation (refining the guesses). Forms of evolutionary computation include swarm intelligence algorithms (such as ant colony or particle swarm optimization)[161] and evolutionary algorithms (such as genetic algorithms, gene expression programming, and genetic programming).[162]

Logic[163] is used for knowledge representation and problem solving, but it can be applied to other problems as well. For example, the satplan algorithm uses logic for planning[164] and inductive logic programming is a method for learning.[165]

Several different forms of logic are used in AI research. Propositional or sentential logic[166] is the logic of statements which can be true or false. First-order logic[167] also allows the use of quantifiers and predicates, and can express facts about objects, their properties, and their relations with each other. Fuzzy logic,[168] is a version of first-order logic which allows the truth of a statement to be represented as a value between 0 and 1, rather than simply True (1) or False (0). Fuzzy systems can be used for uncertain reasoning and have been widely used in modern industrial and consumer product control systems. Subjective logic[citation needed] models uncertainty in a different and more explicit manner than fuzzy-logic: a given binomial opinion satisfies belief + disbelief + uncertainty = 1 within a Beta distribution. By this method, ignorance can be distinguished from probabilistic statements that an agent makes with high confidence.

Default logics, non-monotonic logics and circumscription[88] are forms of logic designed to help with default reasoning and the qualification problem. Several extensions of logic have been designed to handle specific domains of knowledge, such as: description logics;[76] situation calculus, event calculus and fluent calculus (for representing events and time);[77] causal calculus;[78] belief calculus;[169] and modal logics.[79]

Many problems in AI (in reasoning, planning, learning, perception and robotics) require the agent to operate with incomplete or uncertain information. AI researchers have devised a number of powerful tools to solve these problems using methods from probability theory and economics.[170]

Bayesian networks[171] are a very general tool that can be used for a large number of problems: reasoning (using the Bayesian inference algorithm),[172] learning (using the expectation-maximization algorithm),[d][174] planning (using decision networks)[175] and perception (using dynamic Bayesian networks).[176] Bayesian networks are used in AdSense to choose what ads to place and on XBox Live to rate and match players. Probabilistic algorithms can also be used for filtering, prediction, smoothing and finding explanations for streams of data, helping perception systems to analyze processes that occur over time (e.g., hidden Markov models or Kalman filters).[176]

A key concept from the science of economics is “utility”: a measure of how valuable something is to an intelligent agent. Precise mathematical tools have been developed that analyze how an agent can make choices and plan, using decision theory, decision analysis,[178] and information value theory.[94] These tools include models such as Markov decision processes,[179] dynamic decision networks,[176] game theory and mechanism design.[180]

The simplest AI applications can be divided into two types: classifiers (“if shiny then diamond”) and controllers (“if shiny then pick up”). Controllers do, however, also classify conditions before inferring actions, and therefore classification forms a central part of many AI systems. Classifiers are functions that use pattern matching to determine a closest match. They can be tuned according to examples, making them very attractive for use in AI. These examples are known as observations or patterns. In supervised learning, each pattern belongs to a certain predefined class. A class can be seen as a decision that has to be made. All the observations combined with their class labels are known as a data set. When a new observation is received, that observation is classified based on previous experience.[181]

A classifier can be trained in various ways; there are many statistical and machine learning approaches. The decision tree[182] is perhaps the most widely used machine learning algorithm. Other widely used classifiers are the neural network,[184] k-nearest neighbor algorithm,[e][186] kernel methods such as the support vector machine (SVM),[f][188] Gaussian mixture model[189] and the extremely popular naive Bayes classifier.[g][191] The performance of these classifiers have been compared over a wide range of tasks. Classifier performance depends greatly on the characteristics of the data to be classified. There is no single classifier that works best on all given problems; this is also referred to as the “no free lunch” theorem. Determining a suitable classifier for a given problem is still more an art than science.[192]

Neural networks, or neural nets, were inspired by the architecture of neurons in the human brain. A simple “neuron” N accepts input from multiple other neurons, each of which, when activated (or “fired”), cast a weighted “vote” for or against whether neuron N should itself activate. Learning requires an algorithm to adjust these weights based on the training data; one simple algorithm (dubbed “fire together, wire together”) is to increase the weight between two connected neurons when the activation of one triggers the successful activation of another. The net forms “concepts” that are distributed among a subnetwork of shared[h] neurons that tend to fire together; a concept meaning “leg” might be coupled with a subnetwork meaning “foot” that includes the sound for “foot”. Neurons have a continuous spectrum of activation; in addition, neurons can process inputs in a nonlinear way rather than weighing straightforward votes. Modern neural nets can learn both continuous functions and, surprisingly, digital logical operations. Neural networks’ early successes included predicting the stock market and (in 1995) a mostly self-driving car.[i] In the 2010s, advances in neural networks using deep learning thrust AI into widespread public consciousness and contributed to an enormous upshift in corporate AI spending; for example, AI-related M&A in 2017 was over 25 times as large as in 2015.[194][195]

The study of non-learning artificial neural networks[184] began in the decade before the field of AI research was founded, in the work of Walter Pitts and Warren McCullouch. Frank Rosenblatt invented the perceptron, a learning network with a single layer, similar to the old concept of linear regression. Early pioneers also include Alexey Grigorevich Ivakhnenko, Teuvo Kohonen, Stephen Grossberg, Kunihiko Fukushima, Christoph von der Malsburg, David Willshaw, Shun-Ichi Amari, Bernard Widrow, John Hopfield, Eduardo R. Caianiello, and others.

The main categories of networks are acyclic or feedforward neural networks (where the signal passes in only one direction) and recurrent neural networks (which allow feedback and short-term memories of previous input events). Among the most popular feedforward networks are perceptrons, multi-layer perceptrons and radial basis networks.[196] Neural networks can be applied to the problem of intelligent control (for robotics) or learning, using such techniques as Hebbian learning (“fire together, wire together”), GMDH or competitive learning.[197]

Today, neural networks are often trained by the backpropagation algorithm, which had been around since 1970 as the reverse mode of automatic differentiation published by Seppo Linnainmaa,[198][199] and was introduced to neural networks by Paul Werbos.[200][201][202]

Hierarchical temporal memory is an approach that models some of the structural and algorithmic properties of the neocortex.[203]

In short, most neural networks use some form of gradient descent on a hand-created neural topology. However, some research groups, such as Uber, argue that simple neuroevolution to mutate new neural network topologies and weights may be competitive with sophisticated gradient descent approaches. One advantage of neuroevolution is that it may be less prone to get caught in “dead ends”.[204]

Deep learning is any artificial neural network that can learn a long chain of causal links. For example, a feedforward network with six hidden layers can learn a seven-link causal chain (six hidden layers + output layer) and has a “credit assignment path” (CAP) depth of seven. Many deep learning systems need to be able to learn chains ten or more causal links in length.[205] Deep learning has transformed many important subfields of artificial intelligence, including computer vision, speech recognition, natural language processing and others.[206][207][205]

According to one overview,[208] the expression “Deep Learning” was introduced to the Machine Learning community by Rina Dechter in 1986[209] and gained traction after Igor Aizenberg and colleagues introduced it to Artificial Neural Networks in 2000.[210] The first functional Deep Learning networks were published by Alexey Grigorevich Ivakhnenko and V. G. Lapa in 1965.[211][pageneeded] These networks are trained one layer at a time. Ivakhnenko’s 1971 paper[212] describes the learning of a deep feedforward multilayer perceptron with eight layers, already much deeper than many later networks. In 2006, a publication by Geoffrey Hinton and Ruslan Salakhutdinov introduced another way of pre-training many-layered feedforward neural networks (FNNs) one layer at a time, treating each layer in turn as an unsupervised restricted Boltzmann machine, then using supervised backpropagation for fine-tuning. Similar to shallow artificial neural networks, deep neural networks can model complex non-linear relationships. Over the last few years, advances in both machine learning algorithms and computer hardware have led to more efficient methods for training deep neural networks that contain many layers of non-linear hidden units and a very large output layer.[214]

Deep learning often uses convolutional neural networks (CNNs), whose origins can be traced back to the Neocognitron introduced by Kunihiko Fukushima in 1980.[215] In 1989, Yann LeCun and colleagues applied backpropagation to such an architecture. In the early 2000s, in an industrial application CNNs already processed an estimated 10% to 20% of all the checks written in the US.[216] Since 2011, fast implementations of CNNs on GPUs have won many visual pattern recognition competitions.[205]

CNNs with 12 convolutional layers were used in conjunction with reinforcement learning by Deepmind’s “AlphaGo Lee”, the program that beat a top Go champion in 2016.[217]

Early on, deep learning was also applied to sequence learning with recurrent neural networks (RNNs)[218] which are in theory Turing complete[219] and can run arbitrary programs to process arbitrary sequences of inputs. The depth of an RNN is unlimited and depends on the length of its input sequence; thus, an RNN is an example of deep learning.[205] RNNs can be trained by gradient descent[220][221][222] but suffer from the vanishing gradient problem.[206][223] In 1992, it was shown that unsupervised pre-training of a stack of recurrent neural networks can speed up subsequent supervised learning of deep sequential problems.[224]

Numerous researchers now use variants of a deep learning recurrent NN called the long short-term memory (LSTM) network published by Hochreiter & Schmidhuber in 1997.[225] LSTM is often trained by Connectionist Temporal Classification (CTC).[226] At Google, Microsoft and Baidu this approach has revolutionised speech recognition.[227][228][229] For example, in 2015, Google’s speech recognition experienced a dramatic performance jump of 49% through CTC-trained LSTM, which is now available through Google Voice to billions of smartphone users.[230] Google also used LSTM to improve machine translation,[231] Language Modeling[232] and Multilingual Language Processing.[233] LSTM combined with CNNs also improved automatic image captioning[234] and a plethora of other applications.

Early symbolic AI inspired Lisp[235] and Prolog,[236] which dominated early AI programming. Modern AI development often uses mainstream languages such as Python or C++,[237] or niche languages such as Wolfram Language.[238]

In 1950, Alan Turing proposed a general procedure to test the intelligence of an agent now known as the Turing test. This procedure allows almost all the major problems of artificial intelligence to be tested. However, it is a very difficult challenge and at present all agents fail.[239]

Artificial intelligence can also be evaluated on specific problems such as small problems in chemistry, hand-writing recognition and game-playing. Such tests have been termed subject matter expert Turing tests. Smaller problems provide more achievable goals and there are an ever-increasing number of positive results.[citation needed]

For example, performance at draughts (i.e. checkers) is optimal,[citation needed] performance at chess is high-human and nearing super-human (see computer chess:computers versus human) and performance at many everyday tasks (such as recognizing a face or crossing a room without bumping into something) is sub-human.

A quite different approach measures machine intelligence through tests which are developed from mathematical definitions of intelligence. Examples of these kinds of tests start in the late nineties devising intelligence tests using notions from Kolmogorov complexity and data compression.[240] Two major advantages of mathematical definitions are their applicability to nonhuman intelligences and their absence of a requirement for human testers.

A derivative of the Turing test is the Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA). As the name implies, this helps to determine that a user is an actual person and not a computer posing as a human. In contrast to the standard Turing test, CAPTCHA is administered by a machine and targeted to a human as opposed to being administered by a human and targeted to a machine. A computer asks a user to complete a simple test then generates a grade for that test. Computers are unable to solve the problem, so correct solutions are deemed to be the result of a person taking the test. A common type of CAPTCHA is the test that requires the typing of distorted letters, numbers or symbols that appear in an image undecipherable by a computer.

AI is relevant to any intellectual task. Modern artificial intelligence techniques are pervasive and are too numerous to list here. Frequently, when a technique reaches mainstream use, it is no longer considered artificial intelligence; this phenomenon is described as the AI effect.

High-profile examples of AI include autonomous vehicles (such as drones and self-driving cars), medical diagnosis, creating art (such as poetry), proving mathematical theorems, playing games (such as Chess or Go), search engines (such as Google search), online assistants (such as Siri), image recognition in photographs, spam filtering, prediction of judicial decisions[244] and targeting online advertisements.[245][246]

With social media sites overtaking TV as a source for news for young people and news organisations increasingly reliant on social media platforms for generating distribution,[247] major publishers now use artificial intelligence (AI) technology to post stories more effectively and generate higher volumes of traffic.[248]

There are a number of competitions and prizes to promote research in artificial intelligence. The main areas promoted are: general machine intelligence, conversational behavior, data-mining, robotic cars, robot soccer and games.

Artificial intelligence is breaking into the healthcare industry by assisting doctors. According to Bloomberg Technology, Microsoft has developed AI to help doctors find the right treatments for cancer.[249] There is a great amount of research and drugs developed relating to cancer. In detail, there are more than 800 medicines and vaccines to treat cancer. This negatively affects the doctors, because there are too many options to choose from, making it more difficult to choose the right drugs for the patients. Microsoft is working on a project to develop a machine called “Hanover”. Its goal is to memorize all the papers necessary to cancer and help predict which combinations of drugs will be most effective for each patient. One project that is being worked on at the moment is fighting myeloid leukemia, a fatal cancer where the treatment has not improved in decades. Another study was reported to have found that artificial intelligence was as good as trained doctors in identifying skin cancers.[250] Another study is using artificial intelligence to try and monitor multiple high-risk patients, and this is done by asking each patient numerous questions based on data acquired from live doctor to patient interactions.[251]

According to CNN, there was a recent study by surgeons at the Children’s National Medical Center in Washington which successfully demonstrated surgery with an autonomous robot. The team supervised the robot while it performed soft-tissue surgery, stitching together a pig’s bowel during open surgery, and doing so better than a human surgeon, the team claimed.[252] IBM has created its own artificial intelligence computer, the IBM Watson, which has beaten human intelligence (at some levels). Watson not only won at the game show Jeopardy! against former champions,[253] but, was declared a hero after successfully diagnosing a women who was suffering from leukemia.[254]

Advancements in AI have contributed to the growth of the automotive industry through the creation and evolution of self-driving vehicles. As of 2016, there are over 30 companies utilizing AI into the creation of driverless cars. A few companies involved with AI include Tesla, Google, and Apple.[255]

Many components contribute to the functioning of self-driving cars. These vehicles incorporate systems such as braking, lane changing, collision prevention, navigation and mapping. Together, these systems, as well as high performance computers, are integrated into one complex vehicle.[256]

Recent developments in autonomous automobiles have made the innovation of self-driving trucks possible, though they are still in the testing phase. The UK government has passed legislation to begin testing of self-driving truck platoons in 2018.[257] Self-driving truck platoons are a fleet of self-driving trucks following the lead of one non-self-driving truck, so the truck platoons aren’t entirely autonomous yet. Meanwhile, the Daimler, a German automobile corporation, is testing the Freightliner Inspiration which is a semi-autonomous truck that will only be used on the highway.[258]

One main factor that influences the ability for a driver-less automobile to function is mapping. In general, the vehicle would be pre-programmed with a map of the area being driven. This map would include data on the approximations of street light and curb heights in order for the vehicle to be aware of its surroundings. However, Google has been working on an algorithm with the purpose of eliminating the need for pre-programmed maps and instead, creating a device that would be able to adjust to a variety of new surroundings.[259] Some self-driving cars are not equipped with steering wheels or brake pedals, so there has also been research focused on creating an algorithm that is capable of maintaining a safe environment for the passengers in the vehicle through awareness of speed and driving conditions.[260]

Another factor that is influencing the ability for a driver-less automobile is the safety of the passenger. To make a driver-less automobile, engineers must program it to handle high risk situations. These situations could include a head on collision with pedestrians. The car’s main goal should be to make a decision that would avoid hitting the pedestrians and saving the passengers in the car. But there is a possibility the car would need to make a decision that would put someone in danger. In other words, the car would need to decide to save the pedestrians or the passengers.[261] The programing of the car in these situations is crucial to a successful driver-less automobile.

Financial institutions have long used artificial neural network systems to detect charges or claims outside of the norm, flagging these for human investigation. The use of AI in banking can be traced back to 1987 when Security Pacific National Bank in US set-up a Fraud Prevention Task force to counter the unauthorised use of debit cards. Programs like Kasisto and Moneystream are using AI in financial services.

Banks use artificial intelligence systems today to organize operations, maintain book-keeping, invest in stocks, and manage properties. AI can react to changes overnight or when business is not taking place.[262] In August 2001, robots beat humans in a simulated financial trading competition.[263] AI has also reduced fraud and financial crimes by monitoring behavioral patterns of users for any abnormal changes or anomalies.[264]

The use of AI machines in the market in applications such as online trading and decision making has changed major economic theories.[265] For example, AI based buying and selling platforms have changed the law of supply and demand in that it is now possible to easily estimate individualized demand and supply curves and thus individualized pricing. Furthermore, AI machines reduce information asymmetry in the market and thus making markets more efficient while reducing the volume of trades. Furthermore, AI in the markets limits the consequences of behavior in the markets again making markets more efficient. Other theories where AI has had impact include in rational choice, rational expectations, game theory, Lewis turning point, portfolio optimization and counterfactual thinking.

In video games, artificial intelligence is routinely used to generate dynamic purposeful behavior in non-player characters (NPCs). In addition, well-understood AI techniques are routinely used for pathfinding. Some researchers consider NPC AI in games to be a “solved problem” for most production tasks. Games with more atypical AI include the AI director of Left 4 Dead (2008) and the neuroevolutionary training of platoons in Supreme Commander 2 (2010).[266][267]

Worldwide annual military spending on robotics rose from 5.1 billion USD in 2010 to 7.5 billion USD in 2015.[268][269] Military drones capable of autonomous action are widely considered a useful asset. In 2017, Vladimir Putin stated that “Whoever becomes the leader in (artificial intelligence) will become the ruler of the world”.[270][271] Many artificial intelligence researchers seek to distance themselves from military applications of AI.[272]

A platform (or “computing platform”) is defined as “some sort of hardware architecture or software framework (including application frameworks), that allows software to run”. As Rodney Brooks pointed out many years ago, it is not just the artificial intelligence software that defines the AI features of the platform, but rather the actual platform itself that affects the AI that results, i.e., there needs to be work in AI problems on real-world platforms rather than in isolation.

A wide variety of platforms has allowed different aspects of AI to develop, ranging from expert systems such as Cyc to deep-learning frameworks to robot platforms such as the Roomba with open interface.[274] Recent advances in deep artificial neural networks and distributed computing have led to a proliferation of software libraries, including Deeplearning4j, TensorFlow, Theano and Torch.

Collective AI is a platform architecture that combines individual AI into a collective entity, in order to achieve global results from individual behaviors.[275][276] With its collective structure, developers can crowdsource information and extend the functionality of existing AI domains on the platform for their own use, as well as continue to create and share new domains and capabilities for the wider community and greater good.[277] As developers continue to contribute, the overall platform grows more intelligent and is able to perform more requests, providing a scalable model for greater communal benefit.[276] Organizations like SoundHound Inc. and the Harvard John A. Paulson School of Engineering and Applied Sciences have used this collaborative AI model.[278][276]

Original post:

Artificial intelligence – Wikipedia

Posted in Ai

What is Artificial Intelligence (AI)? – Definition from …

Artificial intelligence is a branch of computer science that aims to create intelligent machines. It has become an essential part of the technology industry.

Research associated with artificial intelligence is highly technical and specialized. The core problems of artificial intelligence include programming computers for certain traits such as:

Knowledge engineering is a core part of AI research. Machines can often act and react like humans only if they have abundant information relating to the world. Artificial intelligence must have access to objects, categories, properties and relations between all of them to implement knowledge engineering. Initiating common sense, reasoning and problem-solving power in machines is a difficult and tedious task.

Machine learning is also a core part of AI. Learning without any kind of supervision requires an ability to identify patterns in streams of inputs, whereas learning with adequate supervision involves classification and numerical regressions. Classification determines the category an object belongs to and regression deals with obtaining a set of numerical input or output examples, thereby discovering functions enabling the generation of suitable outputs from respective inputs. Mathematical analysis of machine learning algorithms and their performance is a well-defined branch of theoretical computer science often referred to as computational learning theory.

Machine perception deals with the capability to use sensory inputs to deduce the different aspects of the world, while computer vision is the power to analyze visual inputs with a few sub-problems such as facial, object and gesture recognition.

Robotics is also a major field related to AI. Robots require intelligence to handle tasks such as object manipulation and navigation, along with sub-problems of localization, motion planning and mapping.

See original here:

What is Artificial Intelligence (AI)? – Definition from …

Posted in Ai

Adobe Illustrator Artwork – Wikipedia

Adobe Illustrator Artwork (AI) is a proprietary file format developed by Adobe Systems for representing single-page vector-based drawings in either the EPS or PDF formats. The .ai filename extension is used by Adobe Illustrator.

The AI file format was originally a native format called PGF. PDF compatibility is achieved by embedding a complete copy of the PGF data within the saved PDF format file. This format is not related to .pgf using the same name Progressive Graphics Format.[5]

The same dual path approach as for PGF is used when saving EPS-compatible files in recent versions of Illustrator. Early versions of the AI file format are true EPS files with a restricted, compact syntax, with additional semantics represented by Illustrator-specific DSC comments that conform to DSC’s Open Structuring Conventions. These files are identical to their corresponding Illustrator EPS counterparts, but with the EPS procsets (procedure sets) omitted from the file and instead externally referenced using%%Include directives.

Aside from Adobe Illustrator, the following applications can edit .ai files:

Read more:

Adobe Illustrator Artwork – Wikipedia

Posted in Ai

Artificial intelligence – Wikipedia

Artificial intelligence (AI, also machine intelligence, MI) is intelligence demonstrated by machines, in contrast to the natural intelligence (NI) displayed by humans and other animals. In computer science AI research is defined as the study of “intelligent agents”: any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.[1] Colloquially, the term “artificial intelligence” is applied when a machine mimics “cognitive” functions that humans associate with other human minds, such as “learning” and “problem solving”.

The scope of AI is disputed: as machines become increasingly capable, tasks considered as requiring “intelligence” are often removed from the definition, a phenomenon known as the AI effect, leading to the quip, “AI is whatever hasn’t been done yet.”[3] For instance, optical character recognition is frequently excluded from “artificial intelligence”, having become a routine technology.[4] Capabilities generally classified as AI as of 2017[update] include successfully understanding human speech, competing at the highest level in strategic game systems (such as chess and Go[6]), autonomous cars, intelligent routing in content delivery network and military simulations.

Artificial intelligence was founded as an academic discipline in 1956, and in the years since has experienced several waves of optimism,[7][8] followed by disappointment and the loss of funding (known as an “AI winter”),[9][10] followed by new approaches, success and renewed funding.[8][11] For most of its history, AI research has been divided into subfields that often fail to communicate with each other.[12] These sub-fields are based on technical considerations, such as particular goals (e.g. “robotics” or “machine learning”),[13] the use of particular tools (“logic” or “neural networks”), or deep philosophical differences.[14][15][16] Subfields have also been based on social factors (particular institutions or the work of particular researchers).[12]

The traditional problems (or goals) of AI research include reasoning, knowledge representation, planning, learning, natural language processing, perception and the ability to move and manipulate objects.[13] General intelligence is among the field’s long-term goals.[17] Approaches include statistical methods, computational intelligence, and traditional symbolic AI. Many tools are used in AI, including versions of search and mathematical optimization, neural networks and methods based on statistics, probability and economics. The AI field draws upon computer science, mathematics, psychology, linguistics, philosophy and many others.

The field was founded on the claim that human intelligence “can be so precisely described that a machine can be made to simulate it”.[18] This raises philosophical arguments about the nature of the mind and the ethics of creating artificial beings endowed with human-like intelligence, issues which have been explored by myth, fiction and philosophy since antiquity.[19] Some people also consider AI to be a danger to humanity if it progresses unabatedly.[20] Others believe that AI, unlike previous technological revolutions, will create a risk of mass unemployment.[21]

In the twenty-first century, AI techniques have experienced a resurgence following concurrent advances in computer power, large amounts of data, and theoretical understanding; and AI techniques have become an essential part of the technology industry, helping to solve many challenging problems in computer science.[22][11]

Thought-capable artificial beings appeared as storytelling devices in antiquity,[23] and have been common in fiction, as in Mary Shelley’s Frankenstein or Karel apek’s R.U.R. (Rossum’s Universal Robots).[24] These characters and their fates raised many of the same issues now discussed in the ethics of artificial intelligence.[19]

The study of mechanical or “formal” reasoning began with philosophers and mathematicians in antiquity. The study of mathematical logic led directly to Alan Turing’s theory of computation, which suggested that a machine, by shuffling symbols as simple as “0” and “1”, could simulate any conceivable act of mathematical deduction. This insight, that digital computers can simulate any process of formal reasoning, is known as the ChurchTuring thesis.[25] Along with concurrent discoveries in neurobiology, information theory and cybernetics, this led researchers to consider the possibility of building an electronic brain. Turing proposed that “if a human could not distinguish between responses from a machine and a human, the machine could be considered intelligent”.[26] The first work that is now generally recognized as AI was McCullouch and Pitts’ 1943 formal design for Turing-complete “artificial neurons”.

The field of AI research was born at a workshop at Dartmouth College in 1956.[28] Attendees Allen Newell (CMU), Herbert Simon (CMU), John McCarthy (MIT), Marvin Minsky (MIT) and Arthur Samuel (IBM) became the founders and leaders of AI research.[29] They and their students produced programs that the press described as “astonishing”: computers were learning checkers strategies (c. 1954)[31] (and by 1959 were reportedly playing better than the average human),[32] solving word problems in algebra, proving logical theorems (Logic Theorist, first run c. 1956) and speaking English.[33] By the middle of the 1960s, research in the U.S. was heavily funded by the Department of Defense[34] and laboratories had been established around the world.[35] AI’s founders were optimistic about the future: Herbert Simon predicted, “machines will be capable, within twenty years, of doing any work a man can do”. Marvin Minsky agreed, writing, “within a generation… the problem of creating ‘artificial intelligence’ will substantially be solved”.[7]

They failed to recognize the difficulty of some of the remaining tasks. Progress slowed and in 1974, in response to the criticism of Sir James Lighthill and ongoing pressure from the US Congress to fund more productive projects, both the U.S. and British governments cut off exploratory research in AI. The next few years would later be called an “AI winter”,[9] a period when obtaining funding for AI projects was difficult.

In the early 1980s, AI research was revived by the commercial success of expert systems,[37] a form of AI program that simulated the knowledge and analytical skills of human experts. By 1985 the market for AI had reached over a billion dollars. At the same time, Japan’s fifth generation computer project inspired the U.S and British governments to restore funding for academic research.[8] However, beginning with the collapse of the Lisp Machine market in 1987, AI once again fell into disrepute, and a second, longer-lasting hiatus began.[10]

In the late 1990s and early 21st century, AI began to be used for logistics, data mining, medical diagnosis and other areas.[22] The success was due to increasing computational power (see Moore’s law), greater emphasis on solving specific problems, new ties between AI and other fields (such as statistics, economics and mathematics), and a commitment by researchers to mathematical methods and scientific standards.[38] Deep Blue became the first computer chess-playing system to beat a reigning world chess champion, Garry Kasparov on 11 May 1997.

In 2011, a Jeopardy! quiz show exhibition match, IBM’s question answering system, Watson, defeated the two greatest Jeopardy champions, Brad Rutter and Ken Jennings, by a significant margin. Faster computers, algorithmic improvements, and access to large amounts of data enabled advances in machine learning and perception; data-hungry deep learning methods started to dominate accuracy benchmarks around 2012.[41] The Kinect, which provides a 3D bodymotion interface for the Xbox 360 and the Xbox One use algorithms that emerged from lengthy AI research[42] as do intelligent personal assistants in smartphones.[43] In March 2016, AlphaGo won 4 out of 5 games of Go in a match with Go champion Lee Sedol, becoming the first computer Go-playing system to beat a professional Go player without handicaps.[6][44] In the 2017 Future of Go Summit, AlphaGo won a three-game match with Ke Jie,[45] who at the time continuously held the world No. 1 ranking for two years.[46][47] This marked the completion of a significant milestone in the development of Artificial Intelligence as Go is an extremely complex game, more so than Chess.

According to Bloomberg’s Jack Clark, 2015 was a landmark year for artificial intelligence, with the number of software projects that use AI within Google increased from a “sporadic usage” in 2012 to more than 2,700 projects. Clark also presents factual data indicating that error rates in image processing tasks have fallen significantly since 2011.[48] He attributes this to an increase in affordable neural networks, due to a rise in cloud computing infrastructure and to an increase in research tools and datasets.[11] Other cited examples include Microsoft’s development of a Skype system that can automatically translate from one language to another and Facebook’s system that can describe images to blind people.[48]

A typical AI perceives its environment and takes actions that maximize its chance of successfully achieving its goals.[1] An AI’s intended goal function can be simple (“1 if the AI wins a game of Go, 0 otherwise”) or complex (“Do actions mathematically similar to the actions that got you rewards in the past”). Goals can be explicitly defined, or can be induced. If the AI is programmed for “reinforcement learning”, goals can be implicitly induced by rewarding some types of behavior and punishing others.[a] Alternatively, an evolutionary system can induce goals by using a “fitness function” to mutate and preferentially replicate high-scoring AI systems; this is similar to how animals evolved to innately desire certain goals such as finding food, or how dogs can be bred via artificial selection to possess desired traits. Some AI systems, such as nearest-neighbor, instead reason by analogy; these systems are not generally given goals, except to the degree that goals are somehow implicit in their training data. Such systems can still be benchmarked if the non-goal system is framed as a system whose “goal” is to successfully accomplish its narrow classification task.[51]

AI often revolves around the use of algorithms. An algorithm is a set of unambiguous instructions that a mechanical computer can execute.[b] A complex algorithm is often built on top of other, simpler, algorithms. A simple example of an algorithm is the following recipe for optimal play at tic-tac-toe:

Many AI algorithms are capable of learning from data; they can enhance themselves by learning new heuristics (strategies, or “rules of thumb”, that have worked well in the past), or can themselves write other algorithms. Some of the “learners” described below, including Bayesian networks, decision trees, and nearest-neighbor, could theoretically, if given infinite data, time, and memory, learn to approximate any function, including whatever combination of mathematical functions would best describe the entire world. These learners could therefore, in theory, derive all possible knowledge, by considering every possible hypothesis and matching it against the data. In practice, it is almost never possible to consider every possibility, because of the phenomenon of “combinatorial explosion”, where the amount of time needed to solve a problem grows exponentially. Much of AI research involves figuring out how to identify and avoid considering broad swaths of possibililities that are unlikely to be fruitful.[53] For example, when viewing a map and looking for the shortest driving route from Denver to New York in the East, one can in most cases skip looking at any path through San Francisco or other areas far to the West; thus, an AI wielding an pathfinding algorithm like A* can avoid the combinatorial explosion that would ensue if every possible route had to be ponderously considered in turn.[55]

The earliest (and easiest to understand) approach to AI was symbolism (such as formal logic): “If an otherwise healthy adult has a fever, then they may have influenza”. A second, more general, approach is Bayesian inference: “If the current patient has a fever, adjust the probability they have influenza in such-and-such way”. The third major approach, extremely popular in routine business AI applications, is analogizers such as SVM and nearest-neighbor: “After examining the records of known past patients whose temperature, symptoms, age, and other factors mostly match the current patient, X% of those patients turned out to have influenza”. A fourth approach is harder to intuitively understand, but is inspired by how the brain’s machinery works: the neural network approach uses artificial “neurons” that can learn by comparing itself to the desired output and altering the strengths of the connections between its internal neurons to “reinforce” connections that seemed to be useful. These four main approaches can overlap with each other and with evolutionary systems; for example, neural nets can learn to make inferences, to generalize, and to make analogies. Some systems implicitly or explicitly use multiple of these approaches, alongside many other AI and non-AI algorithms; the best approach is often different depending on the problem.[57]

Learning algorithms work on the basis that strategies, algorithms, and inferences that worked well in the past are likely to continue working well in the future. These inferences can be obvious, such as “since the sun rose every morning for the last 10,000 days, it will probably rise tomorrow morning as well”. They can be nuanced, such as “X% of families have geographically separate species with color variants, so there is an Y% chance that undiscovered black swans exist”. Learners also work on the basis of “Occam’s razor”: The simplest theory that explains the data is the likeliest. Therefore, to be successful, a learner must be designed such that it prefers simpler theories to complex theories, except in cases where the complex theory is proven substantially better. Settling on a bad, overly complex theory gerrymandered to fit all the past training data is known as overfitting. Many systems attempt to reduce overfitting by rewarding a theory in accordance with how well it fits the data, but penalizing the theory in accordance with how complex the theory is. Besides classic overfitting, learners can also disappoint by “learning the wrong lesson”. A toy example is that an image classifier trained only on pictures of brown horses and black cats might conclude that all brown patches are likely to be horses. A real-world example is that, unlike humans, current image classifiers don’t determine the spatial relationship between components of the picture; instead, they learn abstract patterns of pixels that humans are oblivious to, but that linearly correlate with images of certain types of real objects. Faintly superimposing such a pattern on a legitimate image results in an “adversarial” image that the system misclassifies.[c][60][61][62]

Compared with humans, existing AI lacks several features of human “commonsense reasoning”; most notably, humans have powerful mechanisms for reasoning about “nave physics” such as space, time, and physical interactions. This enables even young children to easily make inferences like “If I roll this pen off a table, it will fall on the floor”. Humans also have a powerful mechanism of “folk psychology” that helps them to interpret natural-language sentences such as “The city councilmen refused the demonstrators a permit because they advocated violence”. (A generic AI has difficulty inferring whether the councilmen or the demonstrators are the ones alleged to be advocating violence.)[65][66][67] This lack of “common knowledge” means that AI often makes different mistakes than humans make, in ways that can seem incomprehensible. For example, existing self-driving cars cannot reason about the location nor the intentions of pedestrians in the exact way that humans do, and instead must use non-human modes of reasoning to avoid accidents.[68][69][70]

The overall research goal of artificial intelligence is to create technology that allows computers and machines to function in an intelligent manner. The general problem of simulating (or creating) intelligence has been broken down into sub-problems. These consist of particular traits or capabilities that researchers expect an intelligent system to display. The traits described below have received the most attention.[13]

Early researchers developed algorithms that imitated step-by-step reasoning that humans use when they solve puzzles or make logical deductions.[71] By the late 1980s and 1990s, AI research had developed methods for dealing with uncertain or incomplete information, employing concepts from probability and economics.[72]

These algorithms proved to be insufficient for solving large reasoning problems, because they experienced a “combinatorial explosion”: they became exponentially slower as the problems grew larger.[53] In fact, even humans rarely use the step-by-step deduction that early AI research was able to model. They solve most of their problems using fast, intuitive judgements.[73] Modern statistical approaches to AI (e.g. neural networks) mimic this human ability to make a quick guess based on experience, solving many problems as people do. However, they are not capable of step-by-step deduction.

Knowledge representation[74] and knowledge engineering[75] are central to AI research. Many of the problems machines are expected to solve will require extensive knowledge about the world. Among the things that AI needs to represent are: objects, properties, categories and relations between objects;[76] situations, events, states and time;[77] causes and effects;[78] knowledge about knowledge (what we know about what other people know);[79] and many other, less well researched domains. A representation of “what exists” is an ontology: the set of objects, relations, concepts, and properties formally described so that software agents can interpret them. The semantics of these are captured as description logic concepts, roles, and individuals, and typically implemented as classes, properties, and individuals in the Web Ontology Language.[80] The most general ontologies are called upper ontologies, which attempt to provide a foundation for all other knowledge[81] by acting as mediators between domain ontologies that cover specific knowledge about a particular knowledge domain (field of interest or area of concern). Such formal knowledge representations are suitable for content-based indexing and retrieval, scene interpretation, clinical decision support, knowledge discovery via automated reasoning (inferring new statements based on explicitly stated knowledge), etc. Video events are often represented as SWRL rules, which can be used, among others, to automatically generate subtitles for constrained videos.[82]

Among the most difficult problems in knowledge representation are:

Intelligent agents must be able to set goals and achieve them.[89] They need a way to visualize the futurea representation of the state of the world and be able to make predictions about how their actions will change itand be able to make choices that maximize the utility (or “value”) of available choices.[90]

In classical planning problems, the agent can assume that it is the only system acting in the world, allowing the agent to be certain of the consequences of its actions.[91] However, if the agent is not the only actor, then it requires that the agent can reason under uncertainty. This calls for an agent that can not only assess its environment and make predictions, but also evaluate its predictions and adapt based on its assessment.[92]

Multi-agent planning uses the cooperation and competition of many agents to achieve a given goal. Emergent behavior such as this is used by evolutionary algorithms and swarm intelligence.[93]

Machine learning, a fundamental concept of AI research since the field’s inception,[94] is the study of computer algorithms that improve automatically through experience.[95][96]

Unsupervised learning is the ability to find patterns in a stream of input. Supervised learning includes both classification and numerical regression. Classification is used to determine what category something belongs in, after seeing a number of examples of things from several categories. Regression is the attempt to produce a function that describes the relationship between inputs and outputs and predicts how the outputs should change as the inputs change. In reinforcement learning[97] the agent is rewarded for good responses and punished for bad ones. The agent uses this sequence of rewards and punishments to form a strategy for operating in its problem space. These three types of learning can be analyzed in terms of decision theory, using concepts like utility. The mathematical analysis of machine learning algorithms and their performance is a branch of theoretical computer science known as computational learning theory.[citation needed]

Within developmental robotics, developmental learning approaches are elaborated upon to allow robots to accumulate repertoires of novel skills through autonomous self-exploration, social interaction with human teachers, and the use of guidance mechanisms (active learning, maturation, motor synergies, etc.).[98][99]

Natural language processing[102] gives machines the ability to read and understand human language. A sufficiently powerful natural language processing system would enable natural language user interfaces and the acquisition of knowledge directly from human-written sources, such as newswire texts. Some straightforward applications of natural language processing include information retrieval, text mining, question answering[103] and machine translation.[104]

A common method of processing and extracting meaning from natural language is through semantic indexing. Although these indexes require a large volume of user input, it is expected that increases in processor speeds and decreases in data storage costs will result in greater efficiency.

Machine perception[105] is the ability to use input from sensors (such as cameras, microphones, tactile sensors, sonar and others) to deduce aspects of the world. Computer vision[106] is the ability to analyze visual input. A few selected subproblems are speech recognition,[107] facial recognition and object recognition.[108]

The field of robotics[109] is closely related to AI. Intelligence is required for robots to handle tasks such as object manipulation[110] and navigation, with sub-problems such as localization, mapping, and motion planning. These systems require that an agent is able to: Be spatially cognizant of its surroundings, learn from and build a map of its environment, figure out how to get from one point in space to another, and execute that movement (which often involves compliant motion, a process where movement requires maintaining physical contact with an object).[112]

Affective computing is the study and development of systems that can recognize, interpret, process, and simulate human affects. It is an interdisciplinary field spanning computer sciences, psychology, and cognitive science. While the origins of the field may be traced as far back as the early philosophical inquiries into emotion, the more modern branch of computer science originated with Rosalind Picard’s 1995 paper on “affective computing”.[119][120] A motivation for the research is the ability to simulate empathy, where the machine would be able to interpret human emotions and adapts its behavior to give an appropriate response to those emotions.

Emotion and social skills[121] are important to an intelligent agent for two reasons. First, being able to predict the actions of others by understanding their motives and emotional states allow an agent to make better decisions. Concepts such as game theory, decision theory, necessitate that an agent be able to detect and model human emotions. Second, in an effort to facilitate humancomputer interaction, an intelligent machine may want to display emotions (even if it does not experience those emotions itself) to appear more sensitive to the emotional dynamics of human interaction.

A sub-field of AI addresses creativity both theoretically (the philosophical psychological perspective) and practically (the specific implementation of systems that generate novel and useful outputs).

Many researchers think that their work will eventually be incorporated into a machine with artificial general intelligence, combining all the skills mentioned above and even exceeding human ability in most or all these areas.[17][122] A few believe that anthropomorphic features like artificial consciousness or an artificial brain may be required for such a project.[123][124]

Many of the problems above may also require general intelligence, if machines are to solve the problems as well as people do. For example, even specific straightforward tasks, like machine translation, require that a machine read and write in both languages (NLP), follow the author’s argument (reason), know what is being talked about (knowledge), and faithfully reproduce the author’s original intent (social intelligence). A problem like machine translation is considered “AI-complete”, because all of these problems need to be solved simultaneously in order to reach human-level machine performance.

There is no established unifying theory or paradigm that guides AI research. Researchers disagree about many issues.[125] A few of the most long standing questions that have remained unanswered are these: should artificial intelligence simulate natural intelligence by studying psychology or neurobiology? Or is human biology as irrelevant to AI research as bird biology is to aeronautical engineering?[14] Can intelligent behavior be described using simple, elegant principles (such as logic or optimization)? Or does it necessarily require solving a large number of completely unrelated problems?[15] Can intelligence be reproduced using high-level symbols, similar to words and ideas? Or does it require “sub-symbolic” processing?[16] John Haugeland, who coined the term GOFAI (Good Old-Fashioned Artificial Intelligence), also proposed that AI should more properly be referred to as synthetic intelligence,[126] a term which has since been adopted by some non-GOFAI researchers.

Stuart Shapiro divides AI research into three approaches, which he calls computational psychology, computational philosophy, and computer science. Computational psychology is used to make computer programs that mimic human behavior.[129] Computational philosophy, is used to develop an adaptive, free-flowing computer mind.[129] Implementing computer science serves the goal of creating computers that can perform tasks that only people could previously accomplish.[129] Together, the humanesque behavior, mind, and actions make up artificial intelligence.

In the 1940s and 1950s, a number of researchers explored the connection between neurobiology, information theory, and cybernetics. Some of them built machines that used electronic networks to exhibit rudimentary intelligence, such as W. Grey Walter’s turtles and the Johns Hopkins Beast. Many of these researchers gathered for meetings of the Teleological Society at Princeton University and the Ratio Club in England.[130] By 1960, this approach was largely abandoned, although elements of it would be revived in the 1980s.

When access to digital computers became possible in the middle 1950s, AI research began to explore the possibility that human intelligence could be reduced to symbol manipulation. The research was centered in three institutions: Carnegie Mellon University, Stanford and MIT, and each one developed its own style of research. John Haugeland named these approaches to AI “good old fashioned AI” or “GOFAI”.[131] During the 1960s, symbolic approaches had achieved great success at simulating high-level thinking in small demonstration programs. Approaches based on cybernetics or neural networks were abandoned or pushed into the background.[132] Researchers in the 1960s and the 1970s were convinced that symbolic approaches would eventually succeed in creating a machine with artificial general intelligence and considered this the goal of their field.

Economist Herbert Simon and Allen Newell studied human problem-solving skills and attempted to formalize them, and their work laid the foundations of the field of artificial intelligence, as well as cognitive science, operations research and management science. Their research team used the results of psychological experiments to develop programs that simulated the techniques that people used to solve problems. This tradition, centered at Carnegie Mellon University would eventually culminate in the development of the Soar architecture in the middle 1980s.[133][134]

Unlike Newell and Simon, John McCarthy felt that machines did not need to simulate human thought, but should instead try to find the essence of abstract reasoning and problem solving, regardless of whether people used the same algorithms.[14] His laboratory at Stanford (SAIL) focused on using formal logic to solve a wide variety of problems, including knowledge representation, planning and learning.[135] Logic was also the focus of the work at the University of Edinburgh and elsewhere in Europe which led to the development of the programming language Prolog and the science of logic programming.[136]

Researchers at MIT (such as Marvin Minsky and Seymour Papert)[137] found that solving difficult problems in vision and natural language processing required ad-hoc solutions they argued that there was no simple and general principle (like logic) that would capture all the aspects of intelligent behavior. Roger Schank described their “anti-logic” approaches as “scruffy” (as opposed to the “neat” paradigms at CMU and Stanford).[15] Commonsense knowledge bases (such as Doug Lenat’s Cyc) are an example of “scruffy” AI, since they must be built by hand, one complicated concept at a time.[138]

When computers with large memories became available around 1970, researchers from all three traditions began to build knowledge into AI applications.[139] This “knowledge revolution” led to the development and deployment of expert systems (introduced by Edward Feigenbaum), the first truly successful form of AI software.[37] The knowledge revolution was also driven by the realization that enormous amounts of knowledge would be required by many simple AI applications.

By the 1980s progress in symbolic AI seemed to stall and many believed that symbolic systems would never be able to imitate all the processes of human cognition, especially perception, robotics, learning and pattern recognition. A number of researchers began to look into “sub-symbolic” approaches to specific AI problems.[16] Sub-symbolic methods manage to approach intelligence without specific representations of knowledge.

This includes embodied, situated, behavior-based, and nouvelle AI. Researchers from the related field of robotics, such as Rodney Brooks, rejected symbolic AI and focused on the basic engineering problems that would allow robots to move and survive.[140] Their work revived the non-symbolic viewpoint of the early cybernetics researchers of the 1950s and reintroduced the use of control theory in AI. This coincided with the development of the embodied mind thesis in the related field of cognitive science: the idea that aspects of the body (such as movement, perception and visualization) are required for higher intelligence.

Interest in neural networks and “connectionism” was revived by David Rumelhart and others in the middle of the 1980s.[141] Neural networks are an example of soft computing — they are solutions to problems which cannot be solved with complete logical certainty, and where an approximate solution is often sufficient. Other soft computing approaches to AI include fuzzy systems, evolutionary computation and many statistical tools. The application of soft computing to AI is studied collectively by the emerging discipline of computational intelligence.[142]

In the 1990s, AI researchers developed sophisticated mathematical tools to solve specific subproblems. These tools are truly scientific, in the sense that their results are both measurable and verifiable, and they have been responsible for many of AI’s recent successes. The shared mathematical language has also permitted a high level of collaboration with more established fields (like mathematics, economics or operations research). Stuart Russell and Peter Norvig describe this movement as nothing less than a “revolution” and “the victory of the neats”.[38] Critics argue that these techniques (with few exceptions) are too focused on particular problems and have failed to address the long-term goal of general intelligence. There is an ongoing debate about the relevance and validity of statistical approaches in AI, exemplified in part by exchanges between Peter Norvig and Noam Chomsky.

In the course of 60 or so years of research, AI has developed a large number of tools to solve the most difficult problems in computer science. A few of the most general of these methods are discussed below.

Many problems in AI can be solved in theory by intelligently searching through many possible solutions:[150] Reasoning can be reduced to performing a search. For example, logical proof can be viewed as searching for a path that leads from premises to conclusions, where each step is the application of an inference rule.[151] Planning algorithms search through trees of goals and subgoals, attempting to find a path to a target goal, a process called means-ends analysis.[152] Robotics algorithms for moving limbs and grasping objects use local searches in configuration space.[110] Many learning algorithms use search algorithms based on optimization.

Simple exhaustive searches[153] are rarely sufficient for most real world problems: the search space (the number of places to search) quickly grows to astronomical numbers. The result is a search that is too slow or never completes. The solution, for many problems, is to use “heuristics” or “rules of thumb” that prioritize choices in favor of those that are more likely to reach a goal, and to do so in a shorter number of steps. In some search methodologies heuristics can also serve to entirely eliminate some choices that are unlikely to lead to a goal (called “pruning the search tree”). Heuristics supply the program with a “best guess” for the path on which the solution lies.[154] Heuristics limit the search for solutions into a smaller sample size.

A very different kind of search came to prominence in the 1990s, based on the mathematical theory of optimization. For many problems, it is possible to begin the search with some form of a guess and then refine the guess incrementally until no more refinements can be made. These algorithms can be visualized as blind hill climbing: we begin the search at a random point on the landscape, and then, by jumps or steps, we keep moving our guess uphill, until we reach the top. Other optimization algorithms are simulated annealing, beam search and random optimization.[155]

Evolutionary computation uses a form of optimization search. For example, they may begin with a population of organisms (the guesses) and then allow them to mutate and recombine, selecting only the fittest to survive each generation (refining the guesses). Forms of evolutionary computation include swarm intelligence algorithms (such as ant colony or particle swarm optimization)[156] and evolutionary algorithms (such as genetic algorithms, gene expression programming, and genetic programming).[157]

Logic[158] is used for knowledge representation and problem solving, but it can be applied to other problems as well. For example, the satplan algorithm uses logic for planning[159] and inductive logic programming is a method for learning.[160]

Several different forms of logic are used in AI research. Propositional or sentential logic[161] is the logic of statements which can be true or false. First-order logic[162] also allows the use of quantifiers and predicates, and can express facts about objects, their properties, and their relations with each other. Fuzzy logic,[163] is a version of first-order logic which allows the truth of a statement to be represented as a value between 0 and 1, rather than simply True (1) or False (0). Fuzzy systems can be used for uncertain reasoning and have been widely used in modern industrial and consumer product control systems. Subjective logic[citation needed] models uncertainty in a different and more explicit manner than fuzzy-logic: a given binomial opinion satisfies belief + disbelief + uncertainty = 1 within a Beta distribution. By this method, ignorance can be distinguished from probabilistic statements that an agent makes with high confidence.

Default logics, non-monotonic logics and circumscription[84] are forms of logic designed to help with default reasoning and the qualification problem. Several extensions of logic have been designed to handle specific domains of knowledge, such as: description logics;[76] situation calculus, event calculus and fluent calculus (for representing events and time);[77] causal calculus;[78] belief calculus;[164] and modal logics.[79]

Many problems in AI (in reasoning, planning, learning, perception and robotics) require the agent to operate with incomplete or uncertain information. AI researchers have devised a number of powerful tools to solve these problems using methods from probability theory and economics.[165]

Bayesian networks[166] are a very general tool that can be used for a large number of problems: reasoning (using the Bayesian inference algorithm),[167] learning (using the expectation-maximization algorithm),[d][169] planning (using decision networks)[170] and perception (using dynamic Bayesian networks).[171] Bayesian networks are used in AdSense to choose what ads to place and on XBox Live to rate and match players. Probabilistic algorithms can also be used for filtering, prediction, smoothing and finding explanations for streams of data, helping perception systems to analyze processes that occur over time (e.g., hidden Markov models or Kalman filters).[171]

A key concept from the science of economics is “utility”: a measure of how valuable something is to an intelligent agent. Precise mathematical tools have been developed that analyze how an agent can make choices and plan, using decision theory, decision analysis,[173] and information value theory.[90] These tools include models such as Markov decision processes,[174] dynamic decision networks,[171] game theory and mechanism design.[175]

The simplest AI applications can be divided into two types: classifiers (“if shiny then diamond”) and controllers (“if shiny then pick up”). Controllers do, however, also classify conditions before inferring actions, and therefore classification forms a central part of many AI systems. Classifiers are functions that use pattern matching to determine a closest match. They can be tuned according to examples, making them very attractive for use in AI. These examples are known as observations or patterns. In supervised learning, each pattern belongs to a certain predefined class. A class can be seen as a decision that has to be made. All the observations combined with their class labels are known as a data set. When a new observation is received, that observation is classified based on previous experience.[176]

A classifier can be trained in various ways; there are many statistical and machine learning approaches. The decision tree[177] is perhaps the most widely used machine learning algorithm. Other widely used classifiers are the neural network,[179] k-nearest neighbor algorithm,[e][181] kernel methods such as the support vector machine (SVM),[f][183] Gaussian mixture model[184] and the extremely popular naive Bayes classifier.[g][186] The performance of these classifiers have been compared over a wide range of tasks. Classifier performance depends greatly on the characteristics of the data to be classified. There is no single classifier that works best on all given problems; this is also referred to as the “no free lunch” theorem. Determining a suitable classifier for a given problem is still more an art than science.[187]

Neural networks, or neural nets, were inspired by the architecture of neurons in the human brain. A simple “neuron” N accepts input from multiple other neurons, each of which, when activated (or “fired”), cast a weighted “vote” for or against whether neuron N should itself activate. Learning requires an algorithm to adjust these weights based on the training data; one simple algorithm (dubbed “fire together, wire together”) is to increase the weight between two connected neurons when the activation of one triggers the successful activation of another. The net forms “concepts” that are distributed among a subnetwork of shared[h] neurons that tend to fire together; a concept meaning “leg” might be coupled with a subnetwork meaning “foot” that includes the sound for “foot”. Neurons have a continuous spectrum of activation; in addition, neurons can process inputs in a nonlinear way rather than weighing straightforward votes. Modern neural nets can learn both continuous functions and, surprisingly, digital logical operations. Neural networks’ early successes included predicting the stock market and (in 1995) a mostly self-driving car.[i] In the 2010s, advances in neural networks using deep learning thrust AI into widespread public consciousness and contributed to an enormous upshift in corporate AI spending; for example, AI-related M&A in 2017 was over 25 times as large as in 2015.[189][190]

The study of non-learning artificial neural networks[179] began in the decade before the field of AI research was founded, in the work of Walter Pitts and Warren McCullouch. Frank Rosenblatt invented the perceptron, a learning network with a single layer, similar to the old concept of linear regression. Early pioneers also include Alexey Grigorevich Ivakhnenko, Teuvo Kohonen, Stephen Grossberg, Kunihiko Fukushima, Christoph von der Malsburg, David Willshaw, Shun-Ichi Amari, Bernard Widrow, John Hopfield, Eduardo R. Caianiello, and others.

The main categories of networks are acyclic or feedforward neural networks (where the signal passes in only one direction) and recurrent neural networks (which allow feedback and short-term memories of previous input events). Among the most popular feedforward networks are perceptrons, multi-layer perceptrons and radial basis networks.[191] Neural networks can be applied to the problem of intelligent control (for robotics) or learning, using such techniques as Hebbian learning (“fire together, wire together”), GMDH or competitive learning.[192]

Today, neural networks are often trained by the backpropagation algorithm, which had been around since 1970 as the reverse mode of automatic differentiation published by Seppo Linnainmaa,[193][194] and was introduced to neural networks by Paul Werbos.[195][196][197]

Hierarchical temporal memory is an approach that models some of the structural and algorithmic properties of the neocortex.[198]

In short, most neural networks use some form of gradient descent on a hand-created neural topology. However, some research groups, such as Uber, argue that simple neuroevolution to mutate new neural network topologies and weights may be competitive with sophisticated gradient descent approaches. One advantage of neuroevolution is that it may be less prone to get caught in “dead ends”.[199]

Deep learning is any artificial neural network that can learn a long chain of causal links. For example, a feedforward network with six hidden layers can learn a seven-link causal chain (six hidden layers + output layer) and has a “credit assignment path” (CAP) depth of seven. Many deep learning systems need to be able to learn chains ten or more causal links in length.[200] Deep learning has transformed many important subfields of artificial intelligence, including computer vision, speech recognition, natural language processing and others.[201][202][200]

According to one overview,[203] the expression “Deep Learning” was introduced to the Machine Learning community by Rina Dechter in 1986[204] and gained traction after Igor Aizenberg and colleagues introduced it to Artificial Neural Networks in 2000.[205] The first functional Deep Learning networks were published by Alexey Grigorevich Ivakhnenko and V. G. Lapa in 1965.[206][pageneeded] These networks are trained one layer at a time. Ivakhnenko’s 1971 paper[207] describes the learning of a deep feedforward multilayer perceptron with eight layers, already much deeper than many later networks. In 2006, a publication by Geoffrey Hinton and Ruslan Salakhutdinov introduced another way of pre-training many-layered feedforward neural networks (FNNs) one layer at a time, treating each layer in turn as an unsupervised restricted Boltzmann machine, then using supervised backpropagation for fine-tuning. Similar to shallow artificial neural networks, deep neural networks can model complex non-linear relationships. Over the last few years, advances in both machine learning algorithms and computer hardware have led to more efficient methods for training deep neural networks that contain many layers of non-linear hidden units and a very large output layer.[209]

Deep learning often uses convolutional neural networks (CNNs), whose origins can be traced back to the Neocognitron introduced by Kunihiko Fukushima in 1980.[210] In 1989, Yann LeCun and colleagues applied backpropagation to such an architecture. In the early 2000s, in an industrial application CNNs already processed an estimated 10% to 20% of all the checks written in the US.[211] Since 2011, fast implementations of CNNs on GPUs have won many visual pattern recognition competitions.[200]

CNNs with 12 convolutional layers were used in conjunction with reinforcement learning by Deepmind’s “AlphaGo Lee”, the program that beat a top Go champion in 2016.[212]

Early on, deep learning was also applied to sequence learning with recurrent neural networks (RNNs)[213] which are in theory Turing complete[214] and can run arbitrary programs to process arbitrary sequences of inputs. The depth of an RNN is unlimited and depends on the length of its input sequence; thus, an RNN is an example of deep learning.[200] RNNs can be trained by gradient descent[215][216][217] but suffer from the vanishing gradient problem.[201][218] In 1992, it was shown that unsupervised pre-training of a stack of recurrent neural networks can speed up subsequent supervised learning of deep sequential problems.[219]

Numerous researchers now use variants of a deep learning recurrent NN called the long short-term memory (LSTM) network published by Hochreiter & Schmidhuber in 1997.[220] LSTM is often trained by Connectionist Temporal Classification (CTC).[221] At Google, Microsoft and Baidu this approach has revolutionised speech recognition.[222][223][224] For example, in 2015, Google’s speech recognition experienced a dramatic performance jump of 49% through CTC-trained LSTM, which is now available through Google Voice to billions of smartphone users.[225] Google also used LSTM to improve machine translation,[226] Language Modeling[227] and Multilingual Language Processing.[228] LSTM combined with CNNs also improved automatic image captioning[229] and a plethora of other applications.

Early symbolic AI inspired Lisp[230] and Prolog,[231] which dominated early AI programming. Modern AI development often uses mainstream languages such as Python or C++,[232] or niche languages such as Wolfram Language.[233]

In 1950, Alan Turing proposed a general procedure to test the intelligence of an agent now known as the Turing test. This procedure allows almost all the major problems of artificial intelligence to be tested. However, it is a very difficult challenge and at present all agents fail.[234]

Artificial intelligence can also be evaluated on specific problems such as small problems in chemistry, hand-writing recognition and game-playing. Such tests have been termed subject matter expert Turing tests. Smaller problems provide more achievable goals and there are an ever-increasing number of positive results.[citation needed]

For example, performance at draughts (i.e. checkers) is optimal,[citation needed] performance at chess is high-human and nearing super-human (see computer chess:computers versus human) and performance at many everyday tasks (such as recognizing a face or crossing a room without bumping into something) is sub-human.

A quite different approach measures machine intelligence through tests which are developed from mathematical definitions of intelligence. Examples of these kinds of tests start in the late nineties devising intelligence tests using notions from Kolmogorov complexity and data compression.[235] Two major advantages of mathematical definitions are their applicability to nonhuman intelligences and their absence of a requirement for human testers.

A derivative of the Turing test is the Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA). As the name implies, this helps to determine that a user is an actual person and not a computer posing as a human. In contrast to the standard Turing test, CAPTCHA is administered by a machine and targeted to a human as opposed to being administered by a human and targeted to a machine. A computer asks a user to complete a simple test then generates a grade for that test. Computers are unable to solve the problem, so correct solutions are deemed to be the result of a person taking the test. A common type of CAPTCHA is the test that requires the typing of distorted letters, numbers or symbols that appear in an image undecipherable by a computer.

AI is relevant to any intellectual task. Modern artificial intelligence techniques are pervasive and are too numerous to list here. Frequently, when a technique reaches mainstream use, it is no longer considered artificial intelligence; this phenomenon is described as the AI effect.

High-profile examples of AI include autonomous vehicles (such as drones and self-driving cars), medical diagnosis, creating art (such as poetry), proving mathematical theorems, playing games (such as Chess or Go), search engines (such as Google search), online assistants (such as Siri), image recognition in photographs, spam filtering, prediction of judicial decisions[239] and targeting online advertisements.[240][241]

With social media sites overtaking TV as a source for news for young people and news organisations increasingly reliant on social media platforms for generating distribution,[242] major publishers now use artificial intelligence (AI) technology to post stories more effectively and generate higher volumes of traffic.[243]

There are a number of competitions and prizes to promote research in artificial intelligence. The main areas promoted are: general machine intelligence, conversational behavior, data-mining, robotic cars, robot soccer and games.

Artificial intelligence is breaking into the healthcare industry by assisting doctors. According to Bloomberg Technology, Microsoft has developed AI to help doctors find the right treatments for cancer.[244] There is a great amount of research and drugs developed relating to cancer. In detail, there are more than 800 medicines and vaccines to treat cancer. This negatively affects the doctors, because there are too many options to choose from, making it more difficult to choose the right drugs for the patients. Microsoft is working on a project to develop a machine called “Hanover”. Its goal is to memorize all the papers necessary to cancer and help predict which combinations of drugs will be most effective for each patient. One project that is being worked on at the moment is fighting myeloid leukemia, a fatal cancer where the treatment has not improved in decades. Another study was reported to have found that artificial intelligence was as good as trained doctors in identifying skin cancers.[245] Another study is using artificial intelligence to try and monitor multiple high-risk patients, and this is done by asking each patient numerous questions based on data acquired from live doctor to patient interactions.[246]

According to CNN, there was a recent study by surgeons at the Children’s National Medical Center in Washington which successfully demonstrated surgery with an autonomous robot. The team supervised the robot while it performed soft-tissue surgery, stitching together a pig’s bowel during open surgery, and doing so better than a human surgeon, the team claimed.[247] IBM has created its own artificial intelligence computer, the IBM Watson, which has beaten human intelligence (at some levels). Watson not only won at the game show Jeopardy! against former champions,[248] but, was declared a hero after successfully diagnosing a women who was suffering from leukemia.[249]

Advancements in AI have contributed to the growth of the automotive industry through the creation and evolution of self-driving vehicles. As of 2016, there are over 30 companies utilizing AI into the creation of driverless cars. A few companies involved with AI include Tesla, Google, and Apple.[250]

Many components contribute to the functioning of self-driving cars. These vehicles incorporate systems such as braking, lane changing, collision prevention, navigation and mapping. Together, these systems, as well as high performance computers, are integrated into one complex vehicle.[251]

Recent developments in autonomous automobiles have made the innovation of self-driving trucks possible, though they are still in the testing phase. The UK government has passed legislation to begin testing of self-driving truck platoons in 2018.[252] Self-driving truck platoons are a fleet of self-driving trucks following the lead of one non-self-driving truck, so the truck platoons aren’t entirely autonomous yet. Meanwhile, the Daimler, a German automobile corporation, is testing the Freightliner Inspiration which is a semi-autonomous truck that will only be used on the highway.[253]

One main factor that influences the ability for a driver-less automobile to function is mapping. In general, the vehicle would be pre-programmed with a map of the area being driven. This map would include data on the approximations of street light and curb heights in order for the vehicle to be aware of its surroundings. However, Google has been working on an algorithm with the purpose of eliminating the need for pre-programmed maps and instead, creating a device that would be able to adjust to a variety of new surroundings.[254] Some self-driving cars are not equipped with steering wheels or brake pedals, so there has also been research focused on creating an algorithm that is capable of maintaining a safe environment for the passengers in the vehicle through awareness of speed and driving conditions.[255]

Another factor that is influencing the ability for a driver-less automobile is the safety of the passenger. To make a driver-less automobile, engineers must program it to handle high risk situations. These situations could include a head on collision with pedestrians. The car’s main goal should be to make a decision that would avoid hitting the pedestrians and saving the passengers in the car. But there is a possibility the car would need to make a decision that would put someone in danger. In other words, the car would need to decide to save the pedestrians or the passengers.[256] The programing of the car in these situations is crucial to a successful driver-less automobile.

Financial institutions have long used artificial neural network systems to detect charges or claims outside of the norm, flagging these for human investigation. The use of AI in banking can be traced back to 1987 when Security Pacific National Bank in US set-up a Fraud Prevention Task force to counter the unauthorised use of debit cards. Programs like Kasisto and Moneystream are using AI in financial services.

Banks use artificial intelligence systems today to organize operations, maintain book-keeping, invest in stocks, and manage properties. AI can react to changes overnight or when business is not taking place.[257] In August 2001, robots beat humans in a simulated financial trading competition.[258] AI has also reduced fraud and financial crimes by monitoring behavioral patterns of users for any abnormal changes or anomalies.[259]

The use of AI machines in the market in applications such as online trading and decision making has changed major economic theories.[260] For example, AI based buying and selling platforms have changed the law of supply and demand in that it is now possible to easily estimate individualized demand and supply curves and thus individualized pricing. Furthermore, AI machines reduce information asymmetry in the market and thus making markets more efficient while reducing the volume of trades. Furthermore, AI in the markets limits the consequences of behavior in the markets again making markets more efficient. Other theories where AI has had impact include in rational choice, rational expectations, game theory, Lewis turning point, portfolio optimization and counterfactual thinking.

In video games, artificial intelligence is routinely used to generate dynamic purposeful behavior in non-player characters (NPCs). In addition, well-understood AI techniques are routinely used for pathfinding. Some researchers consider NPC AI in games to be a “solved problem” for most production tasks. Games with more atypical AI include the AI director of Left 4 Dead (2008) and the neuroevolutionary training of platoons in Supreme Commander 2 (2010).[261][262]

Worldwide annual military spending on robotics rose from 5.1 billion USD in 2010 to 7.5 billion USD in 2015.[263][264] Military drones capable of autonomous action are widely considered a useful asset. In 2017, Vladimir Putin stated that “Whoever becomes the leader in (artificial intelligence) will become the ruler of the world”.[265][266] Many artificial intelligence researchers seek to distance themselves from military applications of AI.[267]

A platform (or “computing platform”) is defined as “some sort of hardware architecture or software framework (including application frameworks), that allows software to run”. As Rodney Brooks pointed out many years ago, it is not just the artificial intelligence software that defines the AI features of the platform, but rather the actual platform itself that affects the AI that results, i.e., there needs to be work in AI problems on real-world platforms rather than in isolation.

See original here:

Artificial intelligence – Wikipedia

Posted in Ai

Ai Weiwei – Wikipedia

Ai Weiwei (Chinese: ; pinyin: i Wiwi, English pronunciation(helpinfo); born 28 August 1957 in Beijing) is a Chinese contemporary artist and activist. His father’s (Ai Qing) original surname was written Jiang ().[1][2][3] Ai collaborated with Swiss architects Herzog & de Meuron as the artistic consultant on the Beijing National Stadium for the 2008 Olympics.[4] As a political activist, he has been highly and openly critical of the Chinese Government’s stance on democracy and human rights. He has investigated government corruption and cover-ups, in particular the Sichuan schools corruption scandal following the collapse of so-called “tofu-dreg schools” in the 2008 Sichuan earthquake.[5] In 2011, following his arrest at Beijing Capital International Airport on 3 April, he was held for 81 days without any official charges being filed; officials alluded to their allegations of “economic crimes”.[6]

Ai’s father was the Chinese poet Ai Qing,[7] who was denounced during the Anti-Rightist Movement. In 1958, the family was sent to a labour camp in Beidahuang, Heilongjiang, when Ai was one year old. They were subsequently exiled to Shihezi, Xinjiang in 1961, where they lived for 16 years. Upon Mao Zedong’s death and the end of the Cultural Revolution, the family returned to Beijing in 1976.[8]

In 1978, Ai enrolled in the Beijing Film Academy and studied animation.[9] In 1978, he was one of the founders of the early avant garde art group the “Stars”, together with Ma Desheng, Wang Keping, Huang Rui, Li Shuang, Ah Cheng and Qu Leilei. The group disbanded in 1983,[10] yet Ai participated in regular Stars group shows, The Stars: Ten Years, 1989 (Hanart Gallery, Hong Kong and Taipei), and a retrospective exhibition in Beijing in 2007: Origin Point (Today Art Museum, Beijing).[citation needed]

From 1981 to 1993, he lived in the United States. For the first few years, Ai lived in Philadelphia and San Francisco. He studied English at the University of Pennsylvania and the University of California, Berkeley.[12] Later, he moved to New York City.[10] He studied briefly at Parsons School of Design.[13] Ai attended the Art Students League of New York from 1983 to 1986, where he studied with Bruce Dorfman, Knox Martin and Richard Pousette-Dart.[14] He later dropped out of school, and made a living out of drawing street portraits and working odd jobs. During this period, he gained exposure to the works of Marcel Duchamp, Andy Warhol, and Jasper Johns, and began creating conceptual art by altering readymade objects.

Ai befriended beat poet Allen Ginsberg while living in New York, following a chance meeting at a poetry reading where Ginsberg read out several poems about China. Ginsberg had travelled to China and met with Ai’s father, the noted poet Ai Qing, and consequently Ginsberg and Ai became friends.[15]

When he was living in the East Village (from 1983 to 1993), Ai carried a camera with him all the time and would take pictures of his surroundings wherever he was. The resulting collection of photos were later selected and is now known as the New York Photographs.[16]

At the same time, Ai became fascinated by blackjack card games and frequented Atlantic City casinos. He is still regarded in gambling circles as a top tier professional blackjack player according to an article published on blackjackchamp.com.[17][18][19]

In 1993, Ai returned to China after his father became ill.[20] He helped establish the experimental artists’ Beijing East Village and co-published a series of three books about this new generation of artists with Chinese curator Feng Boyi: Black Cover Book (1994), White Cover Book (1995), and Gray Cover Book (1997).[21]

In 1999, Ai moved to Caochangdi, in the northeast of Beijing, and built a studio house his first architectural project. Due to his interest in architecture, he founded the architecture studio FAKE Design, in 2003.[22] In 2000, he co-curated the art exhibition Fuck Off with curator Feng Boyi in Shanghai, China.[23]

Ai is married to artist Lu Qing,[24] and has a son from an extramarital relationship.[25]

In 2005, Ai was invited to start blogging by Sina Weibo, the biggest internet platform in China. He posted his first blog on 19 November. For four years, he “turned out a steady stream of scathing social commentary, criticism of government policy, thoughts on art and architecture, and autobiographical writings.”[26] The blog was shut down by Sina on 28 May 2009. Ai then turned to Twitter and wrote prolifically on the platform, claiming at least eight hours online every day. He wrote almost exclusively in Chinese using the account @aiww.[citation needed] As of 31 December 2013, Ai has declared that he would stop tweeting but the account remains active in forms of retweets and Instagram posts.[27]

Ai supported the Amnesty International petition for Iranian filmmaker Hossein Rajabian and his brother, musician Mehdi Rajabian, and released the news on his Twitter pages.[28]

Ten days after the 8.0-magnitude earthquake in Sichuan province on 12 May 2008, Ai led a team to survey and film the post-quake conditions in various disaster zones. In response to the government’s lack of transparency in revealing names of students who perished in the earthquake due to substandard school campus constructions, Ai recruited volunteers online and launched a “Citizens’ Investigation” to compile names and information of the student victims. On 20 March 2009, he posted a blog titled “Citizens’ Investigation” and wrote: “To remember the departed, to show concern for life, to take responsibility, and for the potential happiness of the survivors, we are initiating a “Citizens’ Investigation.” We will seek out the names of each departed child, and we will remember them.”[29]

As of 14 April 2009, the list had accumulated 5,385 names.[30] Ai published the collected names as well as numerous articles documenting the investigation on his blog which was shut down by Chinese authorities in May 2009.[31] He also posted his list of names of schoolchildren who died on the wall of his office at FAKE Design in Beijing.[32]

Ai suffered headaches and claimed he had difficulty concentrating on his work since returning from Chengdu in August 2009, where he was beaten by the police for trying to testify for Tan Zuoren, a fellow investigator of the shoddy construction and student casualties in the earthquake. On 14 September 2009, Ai was diagnosed to be suffering internal bleeding in a hospital in Munich, Germany, and the doctor arranged for emergency brain surgery.[33] The cerebral hemorrhage is believed to be linked to the police attack.[34][35]

According to the Financial Times, in an attempt to force Ai to leave the country, two accounts used by him had been hacked in a sophisticated attack on Google in China dubbed Operation Aurora, their contents read and copied; his bank accounts were investigated by state security agents who claimed he was under investigation for “unspecified suspected crimes”.[36]

In November 2010, Ai was placed under house arrest by the Chinese police. He said this was to prevent the planned party marking the demolition of his newly built Shanghai studio.[37]

The building was designed and built by Ai upon encouragement and persuasion from a “high official [from Shanghai]” as part of a new cultural area designated by Shanghai Municipal authorities; Ai would have used it as a studio and to teach architecture courses. But now Ai has been accused of erecting the structure without the necessary planning permission and a demolition notice has been ordered, even though, Ai said, officials had been extremely enthusiastic, and the entire application and planning process was “under government supervision”. According to Ai, a number of artists were invited to build new studios in this area of Shanghai because officials wanted to create a cultural area.[38]

On 3 November 2010, Ai said the government had informed him two months earlier that the newly completed studio would be knocked down because it was illegal. Ai complained that this was unfair, as he was “the only one singled out to have my studio destroyed”. The Guardian reported Ai saying Shanghai municipal authorities were “frustrated” by documentaries on subjects they considered sensitive:[38] two of the better known ones featured Shanghai resident Feng Zhenghu, who lived in forced exile for three months in Narita Airport, Tokyo; another well-known documentary focused on Yang Jia, who murdered six Shanghai police officers.[39]

In the end, the party took place without Ai’s presence; his supporters feasted on river crab, an allusion to “harmony”, and a euphemism used to jeer official censorship. Ai was released from house arrest the next day.[40]

Like other activists and intellectuals, Ai was prevented from leaving China in late 2010. Ai suggested that the authorities wanted to prevent him from attending the ceremony in December 2010 to award the 2010 Nobel Peace Prize to fellow dissident Liu Xiaobo.[41] Ai said that he had not been invited to the ceremony, and was attempting to travel to South Korea for a meeting when he was told that he could not leave for reasons of national security.[42]

In the evening of 11 January 2011, Ai’s studio was demolished in a surprise move by the local government.[43][44]

On 3 April 2011, Ai was arrested at Beijing Capital International Airport just before catching a flight to Hong Kong and his studio facilities were searched.[45] A police contingent of approximately 50 officers came to his studio, threw a cordon around it and searched the premises. They took away laptops and the hard drive from the main computer; along with Ai, police also detained eight staff members and Ai’s wife, Lu Qing. Police also visited the mother of Ai’s two-year-old son.[46] While state media originally reported on 6 April that Ai was arrested at the airport because “his departure procedures were incomplete,”[47] the Chinese Ministry of Foreign Affairs said on 7 April that Ai was arrested under investigation for alleged economic crimes.[48] Then, on 8 April, police returned to Ai’s workshop to examine his financial affairs.[49] On 9 April, Ai’s accountant, as well as studio partner Liu Zhenggang and driver Zhang Jingsong, disappeared,[50] while Ai’s assistant Wen Tao has remained missing since Ai’s arrest on 3 April.[51] Ai’s wife said that she was summoned by the Beijing Chaoyang district tax bureau, where she was interrogated about his studio’s tax on 12 April.[52] South China Morning Post reports that Ai received at least two visits from the police, the last being on 31 March three days before his detention apparently with offers of membership to the Chinese People’s Political Consultative Conference. A staff member recalled that Ai had mentioned receiving the offer earlier, “[but Ai] didn’t say if it was a membership of the CPPCC at the municipal or national level, how he responded or whether he accepted it or not.”[52]

On 24 February, amid an online campaign for Middle East-style protests in major Chinese cities by overseas dissidents, Ai posted on his Twitter account: “I didnt care about jasmine at first, but people who are scared by jasmine sent out information about how harmful jasmine is often, which makes me realize that jasmine is what scares them the most. What a jasmine!”[53][54]

Analysts and other activists said Ai had been widely thought to be untouchable, but Nicholas Bequelin from Human Rights Watch suggested that his arrest, calculated to send the message that no one would be immune, must have had the approval of someone in the top leadership.[55] International governments, human rights groups and art institutions, among others, called for Ai’s release, while Chinese officials did not notify Ai’s family of his whereabouts.[56]

State media started describing Ai as a “deviant and a plagiarist” in early 2011.[57] The China Daily subsidiary, the Global Times editorial on 6 April 2011 attacked Ai, saying “Ai Weiwei likes to do something ‘others dare not do.’ He has been close to the red line of Chinese law. Objectively speaking, Chinese society does not have much experience in dealing with such persons. However, as long as Ai Weiwei continuously marches forward, he will inevitably touch the red line one day.”[58] Two days later, the journal scorned Western media for questioning Ai’s charge as a “catch-all crime”, and denounced the use of his political activism as a “legal shield” against everyday crimes. It said “Ai’s detention is one of the many judicial cases handled in China every day. It is pure fantasy to conclude that Ai’s case will be handled specially and unfairly.”[59] Frank Ching expressed in the South China Morning Post that how the Global Times could radically shift its position from one-day to the next was reminiscent of Alice in Wonderland.[60]

Michael Sheridan of The Times suggested that Ai had offered himself to the authorities on a platter with some of his provocative art, particularly photographs of himself nude with only a toy alpaca hiding his modesty with a caption (“grass mud horse covering the middle”). The term possesses a double meaning in Chinese: one possible interpretation was given by Sheridan as: “Fuck your mother, the party central committee”.[61]

Ming Pao in Hong Kong reacted strongly to the state media’s character attack on Ai, saying that authorities had employed “a chain of actions outside the law, doing further damage to an already weak system of laws, and to the overall image of the country.”[57] Pro-Beijing newspaper in Hong Kong, Wen Wei Po, announced that Ai was under arrest for tax evasion, bigamy and spreading indecent images on the internet, and vilified him with multiple instances of strong rhetoric.[62][63] Supporters said “the article should be seen as a mainland media commentary attacking Ai, rather than as an accurate account of the investigation.”[64]

The United States and European Union protested Ai’s detention.[65] The international arts community also mobilised petitions calling for the release of Ai: “1001 Chairs for Ai Weiwei” was organized by Creative Time of New York that calls for artists to bring chairs to Chinese embassies and consulates around the world on 17 April 2011, at 1pm local time “to sit peacefully in support of the artist’s immediate release.”[66][67] Artists in Hong Kong,[68] Germany[68] and Taiwan demonstrated and called for Ai to be released.[69]

One of the major protests by U.S. museums took place on 19 and 20 May when the Museum of Contemporary Art San Diego organized a 24-hour silent protest in which volunteer participants, including community members, media, and museum staff, occupied two traditionally styled Chinese chairs for one-hour periods.[70] The 24-hour sit-in referenced Ai’s sculpture series, Marble Chair, two of which were on view and were subsequently acquired for the Museum’s permanent collection.

The Solomon R. Guggenheim Foundation and the International Council of Museums, which organised petitions, said they had collected more than 90,000 signatures calling for the release of Ai.[71] On 13 April 2011, a group of European intellectuals led by Vclav Havel had issued an open letter to Wen Jiabao, condemning the arrest and demanding the immediate release of Ai. The signatories include Ivan Klma, Ji Grua, Jchym Topol, Elfriede Jelinek, Adam Michnik, Adam Zagajewski, Helmuth Frauendorfer; Bei Ling (Chinese:), a Chinese poet in exile drafted and also signed the open letter.[72]

On 16 May 2011, the Chinese authorities allowed Ai’s wife to visit him briefly. Liu Xiaoyuan, his attorney and personal friend, reported that Wei was in good physical condition and receiving treatment for his chronic diabetes and hypertension; he was not in a prison or hospital but under some form of house arrest.[73]

He is the subject of the 2012 documentary film Ai Weiwei: Never Sorry, directed by American filmmaker Alison Klayman, which received a special jury prize at the 2012 Sundance Film Festival and opened the Hot Docs Canadian International Documentary Festival, North America’s largest documentary festival, in Toronto on 26 April 2012.[74]

On 22 June 2011, the Chinese authorities released Ai from jail after almost three months’ detention on charges of tax evasion.[75] Beijing Fa Ke Cultural Development Ltd. (Chinese: ), a company Ai controlled, had allegedly evaded taxes and intentionally destroyed accounting documents. State media also reports that Ai was granted bail on account of Ai’s “good attitude in confessing his crimes”, willingness to pay back taxes, and his chronic illnesses.[76] According to the Chinese Foreign Ministry, he is prohibited from leaving Beijing without permission for one year.[77][78] Ai’s supporters widely viewed his detention as retaliation for his vocal criticism of the government.[79] On 23 June 2011, professor Wang Yujin of China University of Political Science and Law stated that the release of Ai on bail shows that the Chinese government could not find any solid evidence of Ai’s alleged “economic crime”.[80] On 24 June 2011, Ai told a Radio Free Asia reporter that he was thankful for the support of the Hong Kong public, and praised Hong Kong’s conscious society. Ai also mentioned that his detention by the Chinese regime was hellish (Chinese: ), and stressed that he is forbidden to say too much to reporters.[81]

After his release, his sister gave some details about his detention condition to the press, explaining that he was subjected to a kind of psychological torture: he was detained in a tiny room with constant light, and two guards were set very close to him at all times, and watched him constantly.[82] In November, Chinese authorities were again investigating Ai and his associates, this time under the charge of spreading pornography.[83][84] Lu was subsequently questioned by police, and released after several hours though the exact charges remain unclear.[85][86] In January 2012, in its International Review issue Art in America magazine featured an interview with Ai Weiwei at his home in China. J.J. Camille (the pen name of a Chinese-born writer living in New York), “neither a journalist nor an activist but simply an art lover who wanted to talk to him” had travelled to Beijing the previous September to conduct the interview and to write about his visit to “China’s most famous dissident artist” for the magazine.[87]

On 21 June 2012, Ai’s bail was lifted. Although he is allowed to leave Beijing, the police informed him that he is still prohibited from traveling to other countries because he is “suspected of other crimes,” including pornography, bigamy and illicit exchange of foreign currency.[88][89] Until 2015, he remained under heavy surveillance and restrictions of movement, but continues to criticize through his work.[90][91] In July 2015, he was given a passport and may travel abroad.[92]

In June 2011, the Beijing Local Taxation Bureau demanded a total of over 12 million yuan (US$1.85million) from Beijing Fa Ke Cultural Development Ltd. in unpaid taxes and fines,[93][94] and accorded three days to appeal the demand in writing. According to Ai’s wife, Beijing Fa Ke Cultural Development Ltd. has hired two Beijing lawyers as defense attorneys. Ai’s family state that Ai is “neither the chief executive nor the legal representative of the design company, which is registered in his wife’s name.”

Offers of donations poured in from Ai’s fans across the world when the fine was announced. Eventually an online loan campaign was initiated on 4 November 2011, and close to 9 million RMB was collected within ten days, from 30,000 contributions. Notes were folded into paper planes and thrown over the studio walls, and donations were made in symbolic amounts such as 8964 (4 June 1989, Tiananmen Massacre) or 512 (12 May 2008, Sichuan earthquake). To thank creditors and acknowledge the contributions as loans, Ai designed and issued loan receipts to all who participated in the campaign.[95] Funds raised from the campaign were used as collateral, required by law for an appeal on the tax case. Lawyers acting for Ai submitted an appeal against the fine in January 2012; the Chinese government subsequently agreed to conduct a review.[96]

In June 2012, the court heard the tax appeal case. Ai’s wife, Lu Qing, the legal representative of the design company, attended the hearing. Lu was accompanied by several lawyers and an accountant, but the witnesses they had requested to testify, including Ai, were prevented from attending a court hearing.[97] Ai asserts that the entire matter including the 81 days he spent in jail in 2011 is intended to suppress his provocations. Ai said he had no illusions as to how the case would turn out, as he believes the court will protect the government’s own interests. On 20 June, hundreds of Ai’s supporters gathered outside the Chaoyang District Court in Beijing despite a small army of police officers, some of whom videotaped the crowd and led several people away.[98] On 20 July, Ai’s tax appeal was rejected in court.[99][100] The same day Ai’s studio released “The Fake Case” which tracks the status and history of this case including a timeline and the release of official documents.[101] On 27 September, the court upheld the 2.4million tax evasion fine.[102] Ai had previously deposited 1.33million in a government-controlled account in order to appeal. Ai said he will not pay the remainder because he does not recognize the charge.[103]

In October 2012, authorities revoked the license of Beijing Fa Ke Cultural Development Ltd. for failing to re-register, an annual requirement by the administration. The company was not able to complete this procedure as its materials and stamps were confiscated by the government.[104]

On 26 April 2014, Ai’s name was removed from a group show taking place at the Shanghai Power Station of Art. The exhibition was held to celebrate the fifteenth anniversary of the art prize created by Uli Sigg in 1998, with the purpose of promoting and developing Chinese contemporary art. Ai won the Lifetime Contribution Award in 2008 and was part of the jury during the first three editions of the prize.[105] He was then invited to take part in the group show together with the other selected Chinese artists. Shortly before the exhibition’s opening, some museum workers removed his name from the list of winners and jury members painted on a wall. Also, Ai’s works Sunflower Seeds and Stools were removed from the show and kept in a museum office (see photo on Ai Weiwei’s Instagram).[106] Sigg declared that it was not his decision and that it was a decision of the Power Station of Art and the Shanghai Municipal Bureau of Culture.[105]

In May 2014, the Ullens Center for Contemporary Art, a non-profit art center situated in the 798 art district of Beijing, held a retrospective exhibition in honor of the late curator and scholar, Hans Van Dijk. Ai, a good friend of Hans and a fellow co-founder of the China Art Archives and Warehouse (CAAW), participated in the exhibition with three artworks.[107] On the day of the opening, Ai realized his name was omitted from both Chinese and English versions of the exhibition’s press release. Ai’s assistants went to the art center and removed his works.[108] It is Ai’s belief that, in omitting his name, the museum altered the historical record of van Dijk’s work with him. Ai started his own research about what actually happened, and between 23 and 25 May he interviewed the UCCA’s director, Philip Tinari, the guest curator of the exhibition, Marianne Brouwer, and the UCCA chief, Xue Mei.[107] He published the transcripts of the interviews on Instagram.[109][110][111][112][113][114][115][116][117] In one of the interviews, the CEO of the UCCA, Xue Mei, admitted that, due to the sensitive time of the exhibition, Ai’s name was taken out of the press releases on the day of the opening and it was supposed to be restored afterwards. This was to avoid problems with the Chinese authorities, who threatened to arrest her.[107]

Beijing video works

From 2003 to 2005, Ai Weiwei recorded the results of Beijing’s developing urban infrastructure and its social conditions.

2003, Video, 150 hours

Beginning under the Dabeiyao highway interchange, the vehicle from which Beijing 2003 was shot traveled every road within the Fourth Ring Road of Beijing and documented the road conditions. Approximately 2400 kilometers and 150 hours of footage later, it ended where it began under the Dabeiyao highway interchange. The documentation of these winding alleyways of the city center now largely torn down for redevelopment preserved a visual record of the city that is free of aesthetic judgment.

2004, Video, 10h 13m

Moving from east to west, Chang’an Boulevard traverses Beijing’s most iconic avenue. Along the boulevard’s 45-kilometer length, it recorded the changing densities of its far-flung suburbs, central business districts, and political core. At each 50-meter increment, the artist records a single frame for one minute. The work reveals the rhythm of Beijing as a capital city, its social structure, cityscape, socialist-planned economy, capitalist market, political power center, commercial buildings, and industrial units as pieces of a multi-layered urban collage.

2005, Video, 1h 6m

2005 Video, 1h 50m

Beijing: The Second Ring and Beijing: The Third Ring capture two opposite views of traffic flow on every bridge of each Ring Road, the innermost arterial highways of Beijing. The artist records a single frame for one minute for each view on the bridge. Beijing: The Second Ring was entirely shot on cloudy days, while the segments for Beijing: The Third Ring were entirely shot on sunny days. The films document the historic aspects and modern development of a city with a population of nearly 11 million people.

2007, video, 2h 32m[118]

This video is about Ai Weiwei’s project Fairytale for Europe’s most innovative five-year art event Documenta 12 in Kassel, Germany in 2007: Ai Weiwei invited 1001 Chinese citizens of different ages and from various backgrounds to Germany to experience their own fairytale for 28 days.[119] The 152 minutes film documents the whole process beginning with project preparations, over the challenge that the participants had to face until the actual travel to Germany, as well as the artist’s ideas behind the work. “This is a work I emotionally relate to. It grows and it surprised me” Ai Weiwei in Fairytale.

2008, video, 1h 18m[120]

On 15 December 2008, a citizens’ investigation began with the goal of seeking an explanation for the casualties of the Sichuan earthquake that happened on 12 May 2008. The investigation covered 14 counties and 74 townships within the disaster zone, and studied the conditions of 153 schools that were affected by the earthquake. By gathering and confirming comprehensive details about the students, such as their age, region, school, and grade, the group managed to affirm that there were 5,192 students who perished in the disaster. Among a hundred volunteers, 38 of them participated in fieldwork, with 25 of them being controlled by the Sichuan police for a total of 45 times. This documentary is a structural element of the citizens’ investigation.

2009, looped video, 1h 27m[121]

At 14:28 on 12 May 2008, an 8.0-magnitude earthquake happened in Sichuan, China. Over 5,000 students in primary and secondary schools perished in the earthquake, yet their names went unannounced. In reaction to the government’s lack of transparency, a citizen’s investigation was initiated to find out their names and details about their schools and families. As of 2 September 2009, there were 4,851 confirmed. This video is a tribute to these perished students and a memorial for innocent lives lost.

2009, video, 48m[122]

This video documents the story of Chinese citizen Feng Zhenghu and his struggles to return home. The Shanghai authorities rejected Feng Zhenghu, originated from Wenzhou, Zhejiang, China, from returning to the country for a total of eight times in 2009. On 4 November 2009, Feng Zhenghu attempted to return home for the ninth time but the police from Shanghai used violence and kidnapped him to board a flight to Japan. Feng refused to enter Japan and decided to live in the Immigration Hall at Terminal 1 of the Narita Airport in Tokyo, as an act of protest. He relied on food gifts from tourists for sustenance and lived at a passageway in the Narita Airport for 92 days. He posted updates over Twitter, they attracted much concern and led to wide media coverage from Chinese netizens and international communities. On 31 January, Feng announced an end to his protest at the Narita Airport. On 12 February, Feng was allowed entry to China, where he reunited with his family at home in Shanghai. Ai Weiwei and his assistant Gao Yuan, went from Beijing to interview Feng Zhenghu three times at the Narita Airport of Japan on 16 November 20 November 2009 and 31 January 2010, and documented his life at the airport passageway and the entire process of his return to China. No country should refuse entry to its own citizens.

2009, video, 1h 19m[123]

Ai Weiwei studio production “Laoma Tihua” is a documentary of an incident during Tan Zuoren’s trial on 12 August 2009. Tan Zuoren was charged with “inciting subversion of state power”. Chengdu police detained witnessed during the trial of the civil rights advocate, which is an obstruction of justice and violence. Tan Zuoren was charged as a result of his research and questioning regarding the 5.12 Wenchuan students’ casualties and the corruption resulting poor building construction. Tan Zuoren was sentenced to five years of prison.

2010, video, 3h[124]

In June 2008, Yang Jia carried a knife, a hammer, a gas mask, pepper spray, gloves and Molotov cocktails to the Zhabei Public Security Branch Bureau and killed six police officers, injuring another police officer and a guard. He was arrested on the scene, and was subsequently charged with intentional homicide. In the following six months, while Yang Jia was detained and trials were held, his mother has mysteriously disappeared. This video is a documentary that traces the reasons and motivations behind the tragedy and investigates into a trial process filled with shady cover-ups and questionable decisions. The film provides a glimpse into the realities of a government-controlled judicial system and its impact on the citizens’ lives.

2010, video, 2h 6m[125]

The future dictionary definition of ‘crackdown’ will be: First cover ones head up firmly, and then beat him or her up violently. @aiww In the summer of 2010, the Chinese government began a crackdown on dissent, and Hua Hao Yue Yuan documents the stories of Liu Dejun and Liu Shasha, whose activism and outspoken attitude led them to violent abuse from the authorities. On separate occasions, they were kidnapped, beaten and thrown into remote locations. The incidents attracted much concern over the Internet, as well as wide speculation and theories about what exactly happened. This documentary presents interviews of the two victims, witnesses and concerned netizens. In which it gathers various perspectives about the two beatings, and brings us closer to the brutal reality of Chinas crackdown on crime.

2010, voice recording, 3h 41m[126]

On 24 April 2010 at 00:51, Ai Weiwei (@aiww) started a Twitter campaign to commemorate students who perished in the earthquake in Sichuan on 12 May 2008. 3,444 friends from the Internet delivered voice recordings, the names of 5,205 perished were recited 12,140 times. Remembrance” is an audio work dedicated to the young people who lost their lives in the Sichuan earthquake. It expresses thoughts for the passing of innocent lives and indignation for the cover-ups on truths about sub-standard architecture, which led to the large number of schools that collapsed during the earthquake.

2010, video, 1h 8m[127]

The shooting and editing of this video lasted nearly seven months at the Ai Weiwei studio. It began near the end of 2007 in an interception organized by cat-saving volunteers in Tianjin, and the film locations included Tianjin, Shanghai, Rugao of Jiangsu, Chaoshan of Guangzhou, and Hebei Province. The documentary depicts a complete picture of a chain in the cat-trading industry. Since the end of 2009 when the government began soliciting expert opinion for the Animal Protection Act, the focus of public debate has always been on whether one should be eating cats or not, or whether cat-eating is a Chinese tradition or not. There are even people who would go as far as to say that the call to stop eating cat meat is “imposing the will of the minority on the majority”. Yet the “majority” does not understand the complete truth of cat-meat trading chains: cat theft, cat trafficking, killing cats, selling cats, and eating cats, all the various stages of the trade and how they are distributed across the country, in cities such as Beijing, Tianjin, Shanghai, Nanjing, Suzhou, Wuxi, Rugao, Wuhan, Guangzhou, and Hebei. This well-organized, smooth-running industry chain of cat abuse, cat killing and skinning has already existed among ordinary Chinese folks for 20 years, or perhaps even longer. The degree of civilization of a country can be seen from its attitude towards animals.

2011, video, 1h 1m[128]

This documentary is about the construction project curated by Herzog & de Meuron and Ai Weiwei. One hundred architects from 27 countries were chosen to participate and design a 1000 square meter villa to be built in a new community in Inner Mongolia. The 100 villas would be designed to fit a master plan designed by Ai Weiwei. On 25 January 2008, the 100 architects gathered in Ordos for a first site visit. The film Ordos 100 documents the total of three site visits to Ordos, during which time the master plan and design of each villa was completed. As of 2016, the Ordos 100 project remains unrealized.

2011, video, 54m[129]

As a sequel to Ai Weiwei’s film Lao Ma Ti Hua, the film So Sorry (named after the artist’s 2009 exhibition in Munich, Germany) shows the beginnings of the tension between Ai Weiwei and the Chinese Government. In Lao Ma Ti Hua, Ai Weiwei travels to Chengdu, Sichuan to attend the trial of the civil rights advocate Tan Zuoren, as a witness. In So Sorry, you see the investigation led by Ai Weiwei studio to identify the students who died during the Sichuan earthquake as a result of corruption and poor building constructions leading to the confrontation between Ai Weiwei and the Chengdu police. After being beaten by the police, Ai Weiwei traveled to Munich, Germany to prepare his exhibition at the museum Haus der Kunst. The result of his beating led to intense headaches caused by a brain hemorrhage and was treated by emergency surgery. These events mark the beginning of Ai Weiwei’s struggle and surveillance at the hands of the state police.

2011, video, 2h 22m[130]

This documentary investigates the death of popular Zhaiqiao village leader Qian Yunhui in the fishing village of Yueqing, Zhejiang province. When the local government confiscated marshlands in order to convert them into construction land, the villagers were deprived of the opportunity to cultivate these lands and be fully self-subsistent. Qian Yunhui, unafraid of speaking up for his villagers, travelled to Beijing several times to report this injustice to the central government. In order to silence him, he was detained by local government repeatedly. On 25 December 2010, Qian Yunhui was hit by a truck and died on the scene. News of the incident and photos of the scene quickly spread over the internet. The local government claimed that Qian Yunhui was the victim of an ordinary traffic accident. This film is an investigation conducted by Ai Weiwei studio into the circumstances of the incident and its connection to the land dispute case, mainly based on interviews of family members, villagers and officials. It is an attempt by Ai Weiwei to establish the facts and find out what really happened on 25 December 2010. During shooting and production, Ai Weiwei studio experienced significant obstruction and resistance from local government. The film crew was followed, sometimes physically stopped from shooting certain scenes and there were even attempts to buy off footage. All villagers interviewed for the purposes of this documentary have been interrogated or illegally detained by local government to some extent.

2011, video, 1h 1m[131]

Early in 2008, the district government of Jiading, Shanghai invited Ai Weiwei to build a studio in Malu Township, as a part of the local government’s efforts in developing its cultural assets. By August 2010, the Ai Weiwei Shanghai Studio completed all of its construction work. In October 2010, the Shanghai government declared the Ai Weiwei Shanghai Studio an illegal construction, and was subjected to demolition. On 7 November 2010, when Ai Weiwei was placed under house arrest by public security in Beijing, over 1,000 netizens attended the “River Crab Feast” at the Shanghai Studio. On 11 January 2011, the Shanghai city government forcibly demolished the Ai Weiwei Studio within a day, without any prior notice.

2013, video, 1h 17m[132]

This video tells the story of Liu Ximei, who at her birth in 1985 was given to relatives to be raised because she was born in violation of China’s strict one-child policy. When she was ten years old, Liu was severely injured while working in the fields and lost large amounts of blood. While undergoing treatment at a local hospital, she was given a blood transfusion that was later revealed to be contaminated with HIV. Following this exposure to the virus, Liu contracted AIDS. According to official statistics, in 2001 there were 850,000 AIDS sufferers in China, many of whom contracted the illness in the 1980s and 1990s as the result of a widespread plasma market operating in rural, impoverished areas and using unsafe collection methods.

2014, video, 2h 8m[133]

Ai Weiwei’s Appeal 15,220,910.50 opens with Ai Weiwei’s mother at the Venice Biennial in the summer of 2013 examining Ai’s large S.A.C.R.E.D. installation portraying his 81-day imprisonment. The documentary goes onto chronologically reconstruct the events that occurred from the time he was arrested at the Beijing airport in April 2011 to his final court appeal in September 2012. The film portrays the day-to-day activity surrounding Ai Weiwei, his family and his associates ranging from consistent visits by the authorities, interviews with reporters, support and donations from fans, and court dates. The Film premiered at the International Film Festival Rotterdam on 23 January 2014.

2015, video, 30m[134]

This documentary on the Fukushima Art Project is about artist Ai Weiwei’s investigation of the site as well as the project’s installation process. In August 2014, Ai Weiwei was invited as one of the participating artists for the Fukushima Nuclear Zone by the Japanese art coalition ChimPom, as part of the project Don’t Follow the Wind . Ai accepted the invitation and sent his assistant Ma Yan to the exclusion zone in Japan to investigate the site. The Fukushima Nuclear Exclusion Zone is thus far located within the 20-kilometer radius of land area of the Fukushima Daiichi Nuclear Power Plant. 25,000 people have already been evacuated from the Exclusion Zone. Both water and electric circuits were cut off. Entrance restriction is expected to be relieved in the next thirty years, or even longer. The art project will also be open to public at that time. The three spots usable as exhibition spaces by the artists are all former residential houses, among which exhibition site one and two were used for working and lodging; and exhibition site three was used as a community entertainment facility with an ostrich farm. Ai brought about two projects, “A Ray of Hope” and “Family Album” after analyzing materials and information generated from the site. In “A Ray of Hope”, a solar photovoltaic system is built on exhibition site one, on the second level of the old warehouse. Integral LED lighting devices are used in the two rooms. The lights would turn on automatically from 7 to 10pm, and from 6 to 8am daily. This lighting system is the only light source in the Exclusion Zone after this project was installed. Photos of Ai and his studio staff at Caochangdi that make up project “Family Album” are displayed on exhibition site two and three, in the seven rooms where locals used to live. The twenty-two selected photos are divided in five categories according to types of event spanning eight years. Among these photos, six of them were taken from the site investigation at the 2008 Sichuan earthquake; two were taken during the time when he was illegally detained after pleading the Tan Zuoren case in Chengdu, China in August 2009; and three others taken during his surgical treatment for his head injury from being attacked in the head by police officers in Chengdu; five taken of him being followed by the police and his Beijing studio Fake Design under surveillance due to the studio tax case from 2011 to 2012; four are photos of Ai Weiwei and his family from year 2011 to year 2013; and the other two were taken earlier of him in his studio in Caochangdi (One taken in 2005 and the other in 2006).

A feature documentary directed and co-produced by Ai Weiwei about the global refugee crisis.

Ai’s visual art includes sculptural installations, woodworking, video and photography. “Ai Weiwei: According to What,” adapted and expanded by the Hirshhorn Museum and Sculpture Garden from a 2009 exhibition at Tokyo’s Mori Art Museum, was Ai’s first North American museum retrospective. [135] It opened at the Hirshhorn in Washington, D.C. in 2013, and subsequently traveled to the Brooklyn Museum, New York, [136] and two other venues. His works address his investigation into the aftermath of the Sichuan earthquake and responses to the Chinese government’s detention and surveillance of him. [137] His recent public pieces have called attention to the Syrian refugee crisis.[138]

(1995) Performance in which Ai lets an ancient ceramic urn fall from his hands and smash to pieces on the ground. The performance was memorialized in a series of three photographic still frames.[139]

(2008) Sculpture resembling a park bench or tree trunk, but its cross-section is a map of China. It is four metres long and weighs 635 kilograms. It is made from wood salvaged from Qing Dynasty temples.[140]

(2008) Ming dynasty table cut in half and rejoined at a right angle to rest two feet on the wall and two on the floor. The reconstruction was completed using Chinese period specific joinery techniques.[141]

(20082012) 150 tons of twisted steel reinforcements recovered from the 2008 Sichuan earthquake building collapse sites were straightened out and displayed as an installation.[142]

(2010) Opening in October 2010 at the Tate Museum in London, Ai displayed 100 million handmade and painted porcelain sunflower seeds. These seeds weighed about 150 tons and were made over a span of two and a half years by 1,600 Jingdezhen artisans. This city made porcelain for the government for over one thousand years. The artisans produced the sunflower seeds in the traditional method that the city is known for, in which a thirty step procedure is employed. The sculpture relates back to chairman Mao’s rule and the Chinese Communist Party. The combination of all the seeds represents that together, the people of China can stand up and overthrow the Chinese Communist Party. Along with this, the seeds represent China’s growing mass production stemming from the consumerist culture in the west. The sculpture directly challenges the “Made in China” mantra that China is known for, considering the labor-intensive and traditional method of creating the work.[143]

(2010) Sculptures in marble to resemble the cameras placed in front of Ai’s studio.[144]

(2011) Sculptures of zodiac animals inspired by the water clock-fountain at the Old Summer Palace.[145]

(2014) Han dynasty vase with the Coca-Cola logo brushed on in red acrylic paint.[146]

(2014) 32 Qing dynasty stools joined together in a cluster with legs pointing out.[147]

(2014) Individual porcelain ornaments, each painted with characters for “free speech”, which when set together form a map of China.[148]

(2014) Consisting of 176 2D-portraits in Lego which are set onto a large floor space, Trace was commissioned by the FOR-SITE Foundation, the United States National Park Service and the Golden Gate Park Conservancy. The original installation was at Alcatraz Prison in San Francisco Bay; the 176 portraits being of various political prisoners and prisoners of conscience. After seeing one million visitors during its one-year display at Alcatraz, the installation was moved and put on display at the Hirshhorn Museum in Washington, D.C. (in a modified form; the pieces had to be arranged to fit the circular floor space). The display at the Hirshhorn ran from 28 June 2017 1 January 2018. The display also included two versions of his wallpaper work The Animal That Looks Like a Llama but Is Really an Alpaca and a video running on a loop.[149]

(2017) As the culmination of Ai’s experiences visiting 40 refugee camps in 2016, Law of the Journey featured an all-black, 230-foot-long inflatable boat carrying 258 faceless refugee figures. The art piece is currently on display at the National Gallery in Prague until 7 January 2018.[150]

(2017) Permanent exhibit, unique setting of two Iron Trees from now on frame the Shrine of the Book in Jerusalem, Israel where Dead Sea Scrolls are preserved[151][152]

(2017) On the view in Israel Museum until the end of October 2017, Journey of Laziz is a video installation, showing mental breakdown and overall suffering of tiger living in the “world’s worst ZOO” in Gaza[151][152]

(2017) On view at the Park Avenue Armory through 6 August 2017, Hansel and Gretel is an installation exploring the theme of surveillance. The project, a collaboration of Ai Weiwei and architects Jacques Herzog and Pierre de Meuron, features surveillance cameras equipped with facial recognition software, near-infrared floor projections, tethered, autonomous drones and sonar beacons. A companion website includes a curatorial statement, artist biographies, a livestream of the installation and a timeline of surveillance technology from ancient to modern times.[153]

View original post here:

Ai Weiwei – Wikipedia

Posted in Ai

Wit.ai

Wit.ai makes it easy for developers to build applications and devices that you can talk or text to. Our vision is to empower developers with an open and extensible natural language platform. Wit.ai learns human language from every interaction, and leverages the community: what’s learned is shared across developers.

Read more:

Wit.ai

Posted in Ai

Cleverbot.com – a clever bot – speak to an AI with some …

About Cleverbot

The site Cleverbot.com started in 2006, but the AI was ‘born’ in 1988, when Rollo Carpenter saw how to make his machine learn. It has been learning ever since!

Things you say to Cleverbot today may influence what it says to others in future. The program chooses how to respond to you fuzzily, and contextually, the whole of your conversation being compared to the millions that have taken place before.

Many people say there is no bot – that it is connecting people together, live. The AI can seem human because it says things real people do say, but it is always software, imitating people.

You’ll have seen scissors on Cleverbot. Using them you can share snippets of chats with friends on social networks. Now you can share snips at Cleverbot.com too!

When you sign in to Cleverbot on this blue bar, you can:

Tweak how the AI responds – 3 different ways!Keep a history of multiple conversationsSwitch between conversationsReturn to a conversation on any machinePublish snippets – snips! – for the world to seeFind and follow friendsBe followed yourself!Rate snips, and see the funniest of themReply to snips posted by othersVote on replies, from awful to great!Choose not to show the scissors

Read more:

Cleverbot.com – a clever bot – speak to an AI with some …

Posted in Ai

AI File Extension – What is a .ai file and how do I open it?

An AI file is a drawing created with Adobe Illustrator, a vector graphics editing program. It is composed of paths connected by points, rather than bitmap image data. AI files are commonly used for logos and print media.

AI file open in Adobe Illustrator CC 2017

Since Illustrator image files are saved in a vector format, they can be enlarged without losing any image quality. Some third-party programs can open AI files, but they may rasterize the image, meaning the vector data will be converted to a bitmap format.

To open an Illustrator document in Photoshop, the file must first have PDF Content saved within the file. If it does not contain the PDF Content, then the graphic cannot be opened and will display a default message, stating, “This is an Adobe Illustrator file that was saved without PDF Content. To place or open this file in other applications, it should be re-saved from Adobe Illustrator with the “Create PDF Compatible File” option turned on.”

See the original post here:

AI File Extension – What is a .ai file and how do I open it?

Posted in Ai

Ai Weiwei – Wikipedia

Ai Weiwei (Chinese: ; pinyin: i Wiwi, English pronunciation(helpinfo); born 28 August 1957 in Beijing) is a Chinese contemporary artist and activist. His father’s (Ai Qing) original surname was written Jiang ().[1][2][3] Ai collaborated with Swiss architects Herzog & de Meuron as the artistic consultant on the Beijing National Stadium for the 2008 Olympics.[4] As a political activist, he has been highly and openly critical of the Chinese Government’s stance on democracy and human rights. He has investigated government corruption and cover-ups, in particular the Sichuan schools corruption scandal following the collapse of so-called “tofu-dreg schools” in the 2008 Sichuan earthquake.[5] In 2011, following his arrest at Beijing Capital International Airport on 3 April, he was held for 81 days without any official charges being filed; officials alluded to their allegations of “economic crimes”.[6]

Ai’s father was the Chinese poet Ai Qing,[7] who was denounced during the Anti-Rightist Movement. In 1958, the family was sent to a labour camp in Beidahuang, Heilongjiang, when Ai was one year old. They were subsequently exiled to Shihezi, Xinjiang in 1961, where they lived for 16 years. Upon Mao Zedong’s death and the end of the Cultural Revolution, the family returned to Beijing in 1976.[8]

In 1978, Ai enrolled in the Beijing Film Academy and studied animation.[9] In 1978, he was one of the founders of the early avant garde art group the “Stars”, together with Ma Desheng, Wang Keping, Huang Rui, Li Shuang, Ah Cheng and Qu Leilei. The group disbanded in 1983,[10] yet Ai participated in regular Stars group shows, The Stars: Ten Years, 1989 (Hanart Gallery, Hong Kong and Taipei), and a retrospective exhibition in Beijing in 2007: Origin Point (Today Art Museum, Beijing).[citation needed]

From 1981 to 1993, he lived in the United States. For the first few years, Ai lived in Philadelphia and San Francisco. He studied English at the University of Pennsylvania and the University of California, Berkeley.[12] Later, he moved to New York City.[10] He studied briefly at Parsons School of Design.[13] Ai attended the Art Students League of New York from 1983 to 1986, where he studied with Bruce Dorfman, Knox Martin and Richard Pousette-Dart.[14] He later dropped out of school, and made a living out of drawing street portraits and working odd jobs. During this period, he gained exposure to the works of Marcel Duchamp, Andy Warhol, and Jasper Johns, and began creating conceptual art by altering readymade objects.

Ai befriended beat poet Allen Ginsberg while living in New York, following a chance meeting at a poetry reading where Ginsberg read out several poems about China. Ginsberg had travelled to China and met with Ai’s father, the noted poet Ai Qing, and consequently Ginsberg and Ai became friends.[15]

When he was living in the East Village (from 1983 to 1993), Ai carried a camera with him all the time and would take pictures of his surroundings wherever he was. The resulting collection of photos were later selected and is now known as the New York Photographs.[16]

At the same time, Ai became fascinated by blackjack card games and frequented Atlantic City casinos. He is still regarded in gambling circles as a top tier professional blackjack player according to an article published on blackjackchamp.com.[17][18][19]

In 1993, Ai returned to China after his father became ill.[20] He helped establish the experimental artists’ Beijing East Village and co-published a series of three books about this new generation of artists with Chinese curator Feng Boyi: Black Cover Book (1994), White Cover Book (1995), and Gray Cover Book (1997).[21]

In 1999, Ai moved to Caochangdi, in the northeast of Beijing, and built a studio house his first architectural project. Due to his interest in architecture, he founded the architecture studio FAKE Design, in 2003.[22] In 2000, he co-curated the art exhibition Fuck Off with curator Feng Boyi in Shanghai, China.[23]

Ai is married to artist Lu Qing,[24] and has a son from an extramarital relationship.[25]

In 2005, Ai was invited to start blogging by Sina Weibo, the biggest internet platform in China. He posted his first blog on 19 November. For four years, he “turned out a steady stream of scathing social commentary, criticism of government policy, thoughts on art and architecture, and autobiographical writings.”[26] The blog was shut down by Sina on 28 May 2009. Ai then turned to Twitter and wrote prolifically on the platform, claiming at least eight hours online every day. He wrote almost exclusively in Chinese using the account @aiww.[citation needed] As of 31 December 2013, Ai has declared that he would stop tweeting but the account remains active in forms of retweets and Instagram posts.[27]

Ai supported the Amnesty International petition for Iranian filmmaker Hossein Rajabian and his brother, musician Mehdi Rajabian, and released the news on his Twitter pages.[28]

Ten days after the 8.0-magnitude earthquake in Sichuan province on 12 May 2008, Ai led a team to survey and film the post-quake conditions in various disaster zones. In response to the government’s lack of transparency in revealing names of students who perished in the earthquake due to substandard school campus constructions, Ai recruited volunteers online and launched a “Citizens’ Investigation” to compile names and information of the student victims. On 20 March 2009, he posted a blog titled “Citizens’ Investigation” and wrote: “To remember the departed, to show concern for life, to take responsibility, and for the potential happiness of the survivors, we are initiating a “Citizens’ Investigation.” We will seek out the names of each departed child, and we will remember them.”[29]

As of 14 April 2009, the list had accumulated 5,385 names.[30] Ai published the collected names as well as numerous articles documenting the investigation on his blog which was shut down by Chinese authorities in May 2009.[31] He also posted his list of names of schoolchildren who died on the wall of his office at FAKE Design in Beijing.[32]

Ai suffered headaches and claimed he had difficulty concentrating on his work since returning from Chengdu in August 2009, where he was beaten by the police for trying to testify for Tan Zuoren, a fellow investigator of the shoddy construction and student casualties in the earthquake. On 14 September 2009, Ai was diagnosed to be suffering internal bleeding in a hospital in Munich, Germany, and the doctor arranged for emergency brain surgery.[33] The cerebral hemorrhage is believed to be linked to the police attack.[34][35]

According to the Financial Times, in an attempt to force Ai to leave the country, two accounts used by him had been hacked in a sophisticated attack on Google in China dubbed Operation Aurora, their contents read and copied; his bank accounts were investigated by state security agents who claimed he was under investigation for “unspecified suspected crimes”.[36]

In November 2010, Ai was placed under house arrest by the Chinese police. He said this was to prevent the planned party marking the demolition of his newly built Shanghai studio.[37]

The building was designed and built by Ai upon encouragement and persuasion from a “high official [from Shanghai]” as part of a new cultural area designated by Shanghai Municipal authorities; Ai would have used it as a studio and to teach architecture courses. But now Ai has been accused of erecting the structure without the necessary planning permission and a demolition notice has been ordered, even though, Ai said, officials had been extremely enthusiastic, and the entire application and planning process was “under government supervision”. According to Ai, a number of artists were invited to build new studios in this area of Shanghai because officials wanted to create a cultural area.[38]

On 3 November 2010, Ai said the government had informed him two months earlier that the newly completed studio would be knocked down because it was illegal. Ai complained that this was unfair, as he was “the only one singled out to have my studio destroyed”. The Guardian reported Ai saying Shanghai municipal authorities were “frustrated” by documentaries on subjects they considered sensitive:[38] two of the better known ones featured Shanghai resident Feng Zhenghu, who lived in forced exile for three months in Narita Airport, Tokyo; another well-known documentary focused on Yang Jia, who murdered six Shanghai police officers.[39]

In the end, the party took place without Ai’s presence; his supporters feasted on river crab, an allusion to “harmony”, and a euphemism used to jeer official censorship. Ai was released from house arrest the next day.[40]

Like other activists and intellectuals, Ai was prevented from leaving China in late 2010. Ai suggested that the authorities wanted to prevent him from attending the ceremony in December 2010 to award the 2010 Nobel Peace Prize to fellow dissident Liu Xiaobo.[41] Ai said that he had not been invited to the ceremony, and was attempting to travel to South Korea for a meeting when he was told that he could not leave for reasons of national security.[42]

In the evening of 11 January 2011, Ai’s studio was demolished in a surprise move by the local government.[43][44]

On 3 April 2011, Ai was arrested at Beijing Capital International Airport just before catching a flight to Hong Kong and his studio facilities were searched.[45] A police contingent of approximately 50 officers came to his studio, threw a cordon around it and searched the premises. They took away laptops and the hard drive from the main computer; along with Ai, police also detained eight staff members and Ai’s wife, Lu Qing. Police also visited the mother of Ai’s two-year-old son.[46] While state media originally reported on 6 April that Ai was arrested at the airport because “his departure procedures were incomplete,”[47] the Chinese Ministry of Foreign Affairs said on 7 April that Ai was arrested under investigation for alleged economic crimes.[48] Then, on 8 April, police returned to Ai’s workshop to examine his financial affairs.[49] On 9 April, Ai’s accountant, as well as studio partner Liu Zhenggang and driver Zhang Jingsong, disappeared,[50] while Ai’s assistant Wen Tao has remained missing since Ai’s arrest on 3 April.[51] Ai’s wife said that she was summoned by the Beijing Chaoyang district tax bureau, where she was interrogated about his studio’s tax on 12 April.[52] South China Morning Post reports that Ai received at least two visits from the police, the last being on 31 March three days before his detention apparently with offers of membership to the Chinese People’s Political Consultative Conference. A staff member recalled that Ai had mentioned receiving the offer earlier, “[but Ai] didn’t say if it was a membership of the CPPCC at the municipal or national level, how he responded or whether he accepted it or not.”[52]

On 24 February, amid an online campaign for Middle East-style protests in major Chinese cities by overseas dissidents, Ai posted on his Twitter account: “I didnt care about jasmine at first, but people who are scared by jasmine sent out information about how harmful jasmine is often, which makes me realize that jasmine is what scares them the most. What a jasmine!”[53][54]

Analysts and other activists said Ai had been widely thought to be untouchable, but Nicholas Bequelin from Human Rights Watch suggested that his arrest, calculated to send the message that no one would be immune, must have had the approval of someone in the top leadership.[55] International governments, human rights groups and art institutions, among others, called for Ai’s release, while Chinese officials did not notify Ai’s family of his whereabouts.[56]

State media started describing Ai as a “deviant and a plagiarist” in early 2011.[57] The China Daily subsidiary, the Global Times editorial on 6 April 2011 attacked Ai, saying “Ai Weiwei likes to do something ‘others dare not do.’ He has been close to the red line of Chinese law. Objectively speaking, Chinese society does not have much experience in dealing with such persons. However, as long as Ai Weiwei continuously marches forward, he will inevitably touch the red line one day.”[58] Two days later, the journal scorned Western media for questioning Ai’s charge as a “catch-all crime”, and denounced the use of his political activism as a “legal shield” against everyday crimes. It said “Ai’s detention is one of the many judicial cases handled in China every day. It is pure fantasy to conclude that Ai’s case will be handled specially and unfairly.”[59] Frank Ching expressed in the South China Morning Post that how the Global Times could radically shift its position from one-day to the next was reminiscent of Alice in Wonderland.[60]

Michael Sheridan of The Times suggested that Ai had offered himself to the authorities on a platter with some of his provocative art, particularly photographs of himself nude with only a toy alpaca hiding his modesty with a caption (“grass mud horse covering the middle”). The term possesses a double meaning in Chinese: one possible interpretation was given by Sheridan as: “Fuck your mother, the party central committee”.[61]

Ming Pao in Hong Kong reacted strongly to the state media’s character attack on Ai, saying that authorities had employed “a chain of actions outside the law, doing further damage to an already weak system of laws, and to the overall image of the country.”[57] Pro-Beijing newspaper in Hong Kong, Wen Wei Po, announced that Ai was under arrest for tax evasion, bigamy and spreading indecent images on the internet, and vilified him with multiple instances of strong rhetoric.[62][63] Supporters said “the article should be seen as a mainland media commentary attacking Ai, rather than as an accurate account of the investigation.”[64]

The United States and European Union protested Ai’s detention.[65] The international arts community also mobilised petitions calling for the release of Ai: “1001 Chairs for Ai Weiwei” was organized by Creative Time of New York that calls for artists to bring chairs to Chinese embassies and consulates around the world on 17 April 2011, at 1pm local time “to sit peacefully in support of the artist’s immediate release.”[66][67] Artists in Hong Kong,[68] Germany[68] and Taiwan demonstrated and called for Ai to be released.[69]

One of the major protests by U.S. museums took place on 19 and 20 May when the Museum of Contemporary Art San Diego organized a 24-hour silent protest in which volunteer participants, including community members, media, and museum staff, occupied two traditionally styled Chinese chairs for one-hour periods.[70] The 24-hour sit-in referenced Ai’s sculpture series, Marble Chair, two of which were on view and were subsequently acquired for the Museum’s permanent collection.

The Solomon R. Guggenheim Foundation and the International Council of Museums, which organised petitions, said they had collected more than 90,000 signatures calling for the release of Ai.[71] On 13 April 2011, a group of European intellectuals led by Vclav Havel had issued an open letter to Wen Jiabao, condemning the arrest and demanding the immediate release of Ai. The signatories include Ivan Klma, Ji Grua, Jchym Topol, Elfriede Jelinek, Adam Michnik, Adam Zagajewski, Helmuth Frauendorfer; Bei Ling (Chinese:), a Chinese poet in exile drafted and also signed the open letter.[72]

On 16 May 2011, the Chinese authorities allowed Ai’s wife to visit him briefly. Liu Xiaoyuan, his attorney and personal friend, reported that Wei was in good physical condition and receiving treatment for his chronic diabetes and hypertension; he was not in a prison or hospital but under some form of house arrest.[73]

He is the subject of the 2012 documentary film Ai Weiwei: Never Sorry, directed by American filmmaker Alison Klayman, which received a special jury prize at the 2012 Sundance Film Festival and opened the Hot Docs Canadian International Documentary Festival, North America’s largest documentary festival, in Toronto on 26 April 2012.[74]

On 22 June 2011, the Chinese authorities released Ai from jail after almost three months’ detention on charges of tax evasion.[75] Beijing Fa Ke Cultural Development Ltd. (Chinese: ), a company Ai controlled, had allegedly evaded taxes and intentionally destroyed accounting documents. State media also reports that Ai was granted bail on account of Ai’s “good attitude in confessing his crimes”, willingness to pay back taxes, and his chronic illnesses.[76] According to the Chinese Foreign Ministry, he is prohibited from leaving Beijing without permission for one year.[77][78] Ai’s supporters widely viewed his detention as retaliation for his vocal criticism of the government.[79] On 23 June 2011, professor Wang Yujin of China University of Political Science and Law stated that the release of Ai on bail shows that the Chinese government could not find any solid evidence of Ai’s alleged “economic crime”.[80] On 24 June 2011, Ai told a Radio Free Asia reporter that he was thankful for the support of the Hong Kong public, and praised Hong Kong’s conscious society. Ai also mentioned that his detention by the Chinese regime was hellish (Chinese: ), and stressed that he is forbidden to say too much to reporters.[81]

After his release, his sister gave some details about his detention condition to the press, explaining that he was subjected to a kind of psychological torture: he was detained in a tiny room with constant light, and two guards were set very close to him at all times, and watched him constantly.[82] In November, Chinese authorities were again investigating Ai and his associates, this time under the charge of spreading pornography.[83][84] Lu was subsequently questioned by police, and released after several hours though the exact charges remain unclear.[85][86] In January 2012, in its International Review issue Art in America magazine featured an interview with Ai Weiwei at his home in China. J.J. Camille (the pen name of a Chinese-born writer living in New York), “neither a journalist nor an activist but simply an art lover who wanted to talk to him” had travelled to Beijing the previous September to conduct the interview and to write about his visit to “China’s most famous dissident artist” for the magazine.[87]

On 21 June 2012, Ai’s bail was lifted. Although he is allowed to leave Beijing, the police informed him that he is still prohibited from traveling to other countries because he is “suspected of other crimes,” including pornography, bigamy and illicit exchange of foreign currency.[88][89] Until 2015, he remained under heavy surveillance and restrictions of movement, but continues to criticize through his work.[90][91] In July 2015, he was given a passport and may travel abroad.[92]

In June 2011, the Beijing Local Taxation Bureau demanded a total of over 12 million yuan (US$1.85million) from Beijing Fa Ke Cultural Development Ltd. in unpaid taxes and fines,[93][94] and accorded three days to appeal the demand in writing. According to Ai’s wife, Beijing Fa Ke Cultural Development Ltd. has hired two Beijing lawyers as defense attorneys. Ai’s family state that Ai is “neither the chief executive nor the legal representative of the design company, which is registered in his wife’s name.”

Offers of donations poured in from Ai’s fans across the world when the fine was announced. Eventually an online loan campaign was initiated on 4 November 2011, and close to 9 million RMB was collected within ten days, from 30,000 contributions. Notes were folded into paper planes and thrown over the studio walls, and donations were made in symbolic amounts such as 8964 (4 June 1989, Tiananmen Massacre) or 512 (12 May 2008, Sichuan earthquake). To thank creditors and acknowledge the contributions as loans, Ai designed and issued loan receipts to all who participated in the campaign.[95] Funds raised from the campaign were used as collateral, required by law for an appeal on the tax case. Lawyers acting for Ai submitted an appeal against the fine in January 2012; the Chinese government subsequently agreed to conduct a review.[96]

In June 2012, the court heard the tax appeal case. Ai’s wife, Lu Qing, the legal representative of the design company, attended the hearing. Lu was accompanied by several lawyers and an accountant, but the witnesses they had requested to testify, including Ai, were prevented from attending a court hearing.[97] Ai asserts that the entire matter including the 81 days he spent in jail in 2011 is intended to suppress his provocations. Ai said he had no illusions as to how the case would turn out, as he believes the court will protect the government’s own interests. On 20 June, hundreds of Ai’s supporters gathered outside the Chaoyang District Court in Beijing despite a small army of police officers, some of whom videotaped the crowd and led several people away.[98] On 20 July, Ai’s tax appeal was rejected in court.[99][100] The same day Ai’s studio released “The Fake Case” which tracks the status and history of this case including a timeline and the release of official documents.[101] On 27 September, the court upheld the 2.4million tax evasion fine.[102] Ai had previously deposited 1.33million in a government-controlled account in order to appeal. Ai said he will not pay the remainder because he does not recognize the charge.[103]

In October 2012, authorities revoked the license of Beijing Fa Ke Cultural Development Ltd. for failing to re-register, an annual requirement by the administration. The company was not able to complete this procedure as its materials and stamps were confiscated by the government.[104]

On 26 April 2014, Ai’s name was removed from a group show taking place at the Shanghai Power Station of Art. The exhibition was held to celebrate the fifteenth anniversary of the art prize created by Uli Sigg in 1998, with the purpose of promoting and developing Chinese contemporary art. Ai won the Lifetime Contribution Award in 2008 and was part of the jury during the first three editions of the prize.[105] He was then invited to take part in the group show together with the other selected Chinese artists. Shortly before the exhibition’s opening, some museum workers removed his name from the list of winners and jury members painted on a wall. Also, Ai’s works Sunflower Seeds and Stools were removed from the show and kept in a museum office (see photo on Ai Weiwei’s Instagram).[106] Sigg declared that it was not his decision and that it was a decision of the Power Station of Art and the Shanghai Municipal Bureau of Culture.[105]

In May 2014, the Ullens Center for Contemporary Art, a non-profit art center situated in the 798 art district of Beijing, held a retrospective exhibition in honor of the late curator and scholar, Hans Van Dijk. Ai, a good friend of Hans and a fellow co-founder of the China Art Archives and Warehouse (CAAW), participated in the exhibition with three artworks.[107] On the day of the opening, Ai realized his name was omitted from both Chinese and English versions of the exhibition’s press release. Ai’s assistants went to the art center and removed his works.[108] It is Ai’s belief that, in omitting his name, the museum altered the historical record of van Dijk’s work with him. Ai started his own research about what actually happened, and between 23 and 25 May he interviewed the UCCA’s director, Philip Tinari, the guest curator of the exhibition, Marianne Brouwer, and the UCCA chief, Xue Mei.[107] He published the transcripts of the interviews on Instagram.[109][110][111][112][113][114][115][116][117] In one of the interviews, the CEO of the UCCA, Xue Mei, admitted that, due to the sensitive time of the exhibition, Ai’s name was taken out of the press releases on the day of the opening and it was supposed to be restored afterwards. This was to avoid problems with the Chinese authorities, who threatened to arrest her.[107]

Beijing video works

From 2003 to 2005, Ai Weiwei recorded the results of Beijings developing urban infrastructure and its social conditions.

2003, Video, 150 hours

Beginning under the Dabeiyao highway interchange, the vehicle from which Beijing 2003 was shot traveled every road within the Fourth Ring Road of Beijing and documented the road conditions. Approximately 2400 kilometers and 150 hours of footage later, it ended where it began under the Dabeiyao highway interchange. The documentation of these winding alleyways of the city center now largely torn down for redevelopment preserved a visual record of the city that is free of aesthetic judgment.

2004, Video, 10h 13m

Moving from east to west, Changan Boulevard traverses Beijings most iconic avenue. Along the boulevards 45-kilometer length, it recorded the changing densities of its far-flung suburbs, central business districts, and political core. At each 50-meter increment, the artist records a single frame for one minute. The work reveals the rhythm of Beijing as a capital city, its social structure, cityscape, socialist-planned economy, capitalist market, political power center, commercial buildings, and industrial units as pieces of a multi-layered urban collage.

2005, Video, 1h 6m

2005 Video, 1h 50m

Beijing: The Second Ring and Beijing: The Third Ring capture two opposite views of traffic flow on every bridge of each Ring Road, the innermost arterial highways of Beijing. The artist records a single frame for one minute for each view on the bridge. Beijing: The Second Ring was entirely shot on cloudy days, while the segments for Beijing: The Third Ring were entirely shot on sunny days. The films document the historic aspects and modern development of a city with a population of nearly 11 million people.

2007, video, 2h 32m[118]

This video is about Ai Weiwei’s project Fairytale for Europes most innovative five-year art event Documenta 12 in Kassel, Germany in 2007: Ai Weiwei invited 1001 Chinese citizens of different ages and from various backgrounds to Germany to experience their own fairytale for 28 days.[119] The 152 minutes film documents the whole process beginning with project preparations, over the challenge that the participants had to face until the actual travel to Germany, as well as the artists ideas behind the work. This is a work I emotionally relate to. It grows and it surprised me Ai Weiwei in Fairytale.

2008, video, 1h 18m[120]

On 15 December 2008, a citizens investigation began with the goal of seeking an explanation for the casualties of the Sichuan earthquake that happened on 12 May 2008. The investigation covered 14 counties and 74 townships within the disaster zone, and studied the conditions of 153 schools that were affected by the earthquake. By gathering and confirming comprehensive details about the students, such as their age, region, school, and grade, the group managed to affirm that there were 5,192 students who perished in the disaster. Among a hundred volunteers, 38 of them participated in fieldwork, with 25 of them being controlled by the Sichuan police for a total of 45 times. This documentary is a structural element of the citizens investigation.

2009, looped video, 1h 27m[121]

At 14:28 on 12 May 2008, an 8.0-magnitude earthquake happened in Sichuan, China. Over 5,000 students in primary and secondary schools perished in the earthquake, yet their names went unannounced. In reaction to the governments lack of transparency, a citizens investigation was initiated to find out their names and details about their schools and families. As of 2 September 2009, there were 4,851 confirmed. This video is a tribute to these perished students and a memorial for innocent lives lost.

2009, video, 48m[122]

This video documents the story of Chinese citizen Feng Zhenghu and his struggles to return home. The Shanghai authorities rejected Feng Zhenghu, originated from Wenzhou, Zhejiang, China, from returning to the country for a total of eight times in 2009. On 4 November 2009, Feng Zhenghu attempted to return home for the ninth time but the police from Shanghai used violence and kidnapped him to board a flight to Japan. Feng refused to enter Japan and decided to live in the Immigration Hall at Terminal 1 of the Narita Airport in Tokyo, as an act of protest. He relied on food gifts from tourists for sustenance and lived at a passageway in the Narita Airport for 92 days. He posted updates over Twitter, they attracted much concern and led to wide media coverage from Chinese netizens and international communities. On 31 January, Feng announced an end to his protest at the Narita Airport. On 12 February, Feng was allowed entry to China, where he reunited with his family at home in Shanghai. Ai Weiwei and his assistant Gao Yuan, went from Beijing to interview Feng Zhenghu three times at the Narita Airport of Japan on 16 November 20 November 2009 and 31 January 2010, and documented his life at the airport passageway and the entire process of his return to China. No country should refuse entry to its own citizens.

2009, video, 1h 19m[123]

Ai Weiwei studio production Laoma Tihua is a documentary of an incident during Tan Zuorens trial on 12 August 2009. Tan Zuoren was charged with inciting subversion of state power. Chengdu police detained witnessed during the trial of the civil rights advocate, which is an obstruction of justice and violence. Tan Zuoren was charged as a result of his research and questioning regarding the 5.12 Wenchuan students casualties and the corruption resulting poor building construction. Tan Zuoren was sentenced five years to prison.

2010, video, 3h[124]

In June 2008, Yang Jia carried a knife, a hammer, a gas mask, pepper spray, gloves and Molotov cocktails to the Zhabei Public Security Branch Bureau and killed six police officers, injuring another police officer and a guard. He was arrested on the scene, and was subsequently charged with intentional homicide. In the following six months, while Yang Jia was detained and trials were held, his mother has mysteriously disappeared. This video is a documentary that traces the reasons and motivations behind the tragedy and investigates into a trial process filled with shady cover-ups and questionable decisions. The film provides a glimpse into the realities of a government-controlled judicial system and its impact on the citizens lives.

2010, video, 2h 6m[125]

The future dictionary definition of crackdown will be: First cover ones head up firmly, and then beat him or her up violently. @aiww In the summer of 2010, the Chinese government began a crackdown on dissent, and Hua Hao Yue Yuan documents the stories of Liu Dejun and Liu Shasha, whose activism and outspoken attitude led them to violent abuse from the authorities. On separate occasions, they were kidnapped, beaten and thrown into remote locations. The incidents attracted much concern over the Internet, as well as wide speculation and theories about what exactly happened. This documentary presents interviews of the two victims, witnesses and concerned netizens. In which it gathers various perspectives about the two beatings, and brings us closer to the brutal reality of Chinas crackdown on crime.

2010, voice recording, 3h 41m[126]

On 24 April 2010 at 00:51, Ai Weiwei (@aiww) started a Twitter campaign to commemorate students who perished in the earthquake in Sichuan on 12 May 2008. 3,444 friends from the Internet delivered voice recordings, the names of 5,205 perished were recited 12,140 times. Remembrance is an audio work dedicated to the young people who lost their lives in the Sichuan earthquake. It expresses thoughts for the passing of innocent lives and indignation for the cover-ups on truths about sub-standard architecture, which led to the large number of schools that collapsed during the earthquake.

2010, video, 1h 8m[127]

The shooting and editing of this video lasted nearly seven months at the Ai Weiwei studio. It began near the end of 2007 in an interception organized by cat-saving volunteers in Tianjin, and the film locations included Tianjin, Shanghai, Rugao of Jiangsu, Chaoshan of Guangzhou, and Hebei Province. The documentary depicts a complete picture of a chain in the cat-trading industry. Since the end of 2009 when the government began soliciting expert opinion for the Animal Protection Act, the focus of public debate has always been on whether one should be eating cats or not, or whether cat-eating is a Chinese tradition or not. There are even people who would go as far as to say that the call to stop eating cat meat is “imposing the will of the minority on the majority”. Yet the “majority” does not understand the complete truth of cat-meat trading chains: cat theft, cat trafficking, killing cats, selling cats, and eating cats, all the various stages of the trade and how they are distributed across the country, in cities such as Beijing, Tianjin, Shanghai, Nanjing, Suzhou, Wuxi, Rugao, Wuhan, Guangzhou, and Hebei. This well-organized, smooth-running industry chain of cat abuse, cat killing and skinning has already existed among ordinary Chinese folks for 20 years, or perhaps even longer. The degree of civilization of a country can be seen from its attitude towards animals.

2011, video, 1h 1m[128]

This documentary is about the construction project curated by Herzog & de Meuron and Ai Weiwei. One hundred architects from 27 countries were chosen to participate and design a 1000 square meter villa to be built in a new community in Inner Mongolia. The 100 villas would be designed to fit a master plan designed by Ai Weiwei. On 25 January 2008, the 100 architects gathered in Ordos for a first site visit. The film Ordos 100 documents the total of three site visits to Ordos, during which time the master plan and design of each villa was completed. As of 2016, the Ordos 100 project remains unrealized.

2011, video, 54m[129]

As a sequel to Ai Weiweis film Lao Ma Ti Hua, the film So Sorry (named after the artists 2009 exhibition in Munich, Germany) shows the beginnings of the tension between Ai Weiwei and the Chinese Government. In Lao Ma Ti Hua, Ai Weiwei travels to Chengdu, Sichuan to attend the trial of the civil rights advocate Tan Zuoren, as a witness. In So Sorry, you see the investigation led by Ai Weiwei studio to identify the students who died during the Sichuan earthquake as a result of corruption and poor building constructions leading to the confrontation between Ai Weiwei and the Chengdu police. After being beaten by the police, Ai Weiwei traveled to Munich, Germany to prepare his exhibition at the museum Haus der Kunst. The result of his beating led to intense headaches caused by a brain hemorrhage and was treated by emergency surgery. These events mark the beginning of Ai Weiweis struggle and surveillance at the hands of the state police.

2011, video, 2h 22m[130]

This documentary investigates the death of popular Zhaiqiao village leader Qian Yunhui in the fishing village of Yueqing, Zhejiang province. When the local government confiscated marshlands in order to convert them into construction land, the villagers were deprived of the opportunity to cultivate these lands and be fully self-subsistent. Qian Yunhui, unafraid of speaking up for his villagers, travelled to Beijing several times to report this injustice to the central government. In order to silence him, he was detained by local government repeatedly. On 25 December 2010, Qian Yunhui was hit by a truck and died on the scene. News of the incident and photos of the scene quickly spread over the internet. The local government claimed that Qian Yunhui was the victim of an ordinary traffic accident. This film is an investigation conducted by Ai Weiwei studio into the circumstances of the incident and its connection to the land dispute case, mainly based on interviews of family members, villagers and officials. It is an attempt by Ai Weiwei to establish the facts and find out what really happened on 25 December 2010. During shooting and production, Ai Weiwei studio experienced significant obstruction and resistance from local government. The film crew was followed, sometimes physically stopped from shooting certain scenes and there were even attempts to buy off footage. All villagers interviewed for the purposes of this documentary have been interrogated or illegally detained by local government to some extent.

2011, video, 1h 1m[131]

Early in 2008, the district government of Jiading, Shanghai invited Ai Weiwei to build a studio in Malu Township, as a part of the local government’s efforts in developing its cultural assets. By August 2010, the Ai Weiwei Shanghai Studio completed all of its construction work. In October 2010, the Shanghai government declared the Ai Weiwei Shanghai Studio an illegal construction, and was subjected to demolition. On 7 November 2010, when Ai Weiwei was placed under house arrest by public security in Beijing, over 1,000 netizens attended the “River Crab Feast” at the Shanghai Studio. On 11 January 2011, the Shanghai city government forcibly demolished the Ai Weiwei Studio within a day, without any prior notice.

2013, video, 1h 17m[132]

This video tells the story of Liu Ximei, who at her birth in 1985 was given to relatives to be raised because she was born in violation of Chinas strict one-child policy. When she was ten years old, Liu was severely injured while working in the fields and lost large amounts of blood. While undergoing treatment at a local hospital, she was given a blood transfusion that was later revealed to be contaminated with HIV. Following this exposure to the virus, Liu contracted AIDS. According to official statistics, in 2001 there were 850,000 AIDS sufferers in China, many of whom contracted the illness in the 1980s and 1990s as the result of a widespread plasma market operating in rural, impoverished areas and using unsafe collection methods.

2014, video, 2h 8m[133]

Ai Weiweis Appeal 15,220,910.50 opens with Ai Weiweis mother at the Venice Biennial in the summer of 2013 examining Ais large S.A.C.R.E.D. installation portraying his 81-day imprisonment. The documentary goes onto chronologically reconstruct the events that occurred from the time he was arrested at the Beijing airport in April 2011 to his final court appeal in September 2012. The film portrays the day-to-day activity surrounding Ai Weiwei, his family and his associates ranging from consistent visits by the authorities, interviews with reporters, support and donations from fans, and court dates. The Film premiered at the International Film Festival Rotterdam on 23 January 2014.

2015, video, 30m[134]

This documentary on the Fukushima Art Project is about artist Ai Weiweis investigation of the site as well as the project’s installation process. In August 2014, Ai Weiwei was invited as one of the participating artists for the Fukushima Nuclear Zone by the Japanese art coalition ChimPom, as part of the project Dont Follow the Wind . Ai accepted the invitation and sent his assistant Ma Yan to the exclusion zone in Japan to investigate the site. The Fukushima Nuclear Exclusion Zone is thus far located within the 20-kilometer radius of land area of the Fukushima Daiichi Nuclear Power Plant. 25,000 people have already been evacuated from the Exclusion Zone. Both water and electric circuits were cut off. Entrance restriction is expected to be relieved in the next thirty years, or even longer. The art project will also be open to public at that time. The three spots usable as exhibition spaces by the artists are all former residential houses, among which exhibition site one and two were used for working and lodging; and exhibition site three was used as a community entertainment facility with an ostrich farm. Ai brought about two projects, “A Ray of Hope” and “Family Album” after analyzing materials and information generated from the site. In “A Ray of Hope”, a solar photovoltaic system is built on exhibition site one, on the second level of the old warehouse. Integral LED lighting devices are used in the two rooms. The lights would turn on automatically from 7 to 10pm, and from 6 to 8am daily. This lighting system is the only light source in the Exclusion Zone after this project was installed. Photos of Ai and his studio staff at Caochangdi that make up project “Family Album” are displayed on exhibition site two and three, in the seven rooms where locals used to live. The twenty-two selected photos are divided in five categories according to types of event spanning eight years. Among these photos, six of them were taken from the site investigation at the 2008 Sichuan earthquake; two were taken during the time when he was illegally detained after pleading the Tan Zuoren case in Chengdu, China in August 2009; and three others taken during his surgical treatment for his head injury from being attacked in the head by police officers in Chengdu; five taken of him being followed by the police and his Beijing studio Fake Design under surveillance due to the studio tax case from 2011 to 2012; four are photos of Ai Weiwei and his family from year 2011 to year 2013; and the other two were taken earlier of him in his studio in Caochangdi (One taken in 2005 and the other in 2006).

A feature documentary directed and co-produced by Ai Weiwei about the global refugee crisis.

Ai’s visual art includes sculptural installations, woodworking, video and photography. “Ai Weiwei: According to What,” adapted and expanded by the Hirshhorn Museum and Sculpture Garden from a 2009 exhibition at Tokyo’s Mori Art Museum, was Ai’s first North American museum retrospective. [135] It opened at the Hirshhorn in Washington, D.C. in 2013, and subsequently traveled to the Brooklyn Museum, New York, [136] and two other venues. His works address his investigation into the aftermath of the Sichuan earthquake and responses to the Chinese government’s detention and surveillance of him. [137] His recent public pieces have called attention to the Syrian refugee crisis.[138]

(1995) Performance in which Ai lets an ancient ceramic urn fall from his hands and smash to pieces on the ground. The performance was memorialized in a series of three photographic still frames.[139]

(2008) Sculpture resembling a park bench or tree trunk, but its cross-section is a map of China. It is four metres long and weighs 635 kilograms. It is made from wood salvaged from Qing Dynasty temples.[140]

(2008) Ming dynasty table cut in half and rejoined at a right angle to rest two feet on the wall and two on the floor. The reconstruction was completed using Chinese period specific joinery techniques.[141]

(2008-2012) 150 tons of twisted steel reinforcements recovered from the 2008 Sichuan earthquake building collapse sites were straightened out and displayed as an installation.[142]

(2010) Opening in October 2010 at the Tate Museum in London, Ai displayed 100 million handmade and painted porcelain sunflower seeds. These seeds weight about 150 tons and were made over a span of two and a half years by 1,600 Jingdezhen artisans. This city made porcelain for the government for over one thousand years. The artisans produced the sunflower seeds in the traditional method that the city is known for, in which a thirty step procedure is employed. The sculpture relates back to Chairman Mao’s rule and the Chinese Communist Party. The combination of all the seeds represent that together, the people of China can stand up and overthrow the Chinese Communist Party. Along with this, the seeds represent China’s growing mass production stemming from the consumerist culture in the west. The sculpture directly challenges the Made in China mantra that China is known for, considering the labor-intensive and traditional method of creating the work.[143]

(2010) Sculptures in marble to resemble the cameras placed in front of Ai’s studio.[144]

(2011) Sculptures of zodiac animals inspired by the water clock-fountain at the Old Summer Palace.[145]

(2014) Han dynasty vase with the Coca-Cola logo brushed on in red acrylic paint.[146]

(2014) 32 Qing dynasty stools joined together in a cluster with legs pointing out.[147]

(2014) Individual porcelain ornaments, each painted with characters for “free speech”, which when set together form a map of China.[148]

(2014) Consisting of 176 2D-portraits in Lego which are set onto a large floor space, Trace was commissioned by the FOR-SITE Foundation, the United States National Park Service and the Golden Gate Park Conservancy. The original installation was at Alcatraz Prison in San Francisco Bay; the 176 portraits being of various political prisoners and prisoners of conscience. After seeing one million visitors during its one-year display at Alcatraz, the installation was moved and put on display at the Hirshhorn Museum in Washington, D.C. (in a modified form; the pieces had to be arranged to fit the circular floor space). The display at the Hirshhorn ran from June 28, 2017 January 1, 2018. The display also included two versions of his wallpaper work The Animal That Looks Like a Llama but Is Really an Alpaca and a video running on a loop.[149]

(2017) As the culmination of Ai’s experiences visiting 40 refugee camps in 2016, Law of the Journey featured an all-black, 230-foot-long inflatable boat carrying 258 faceless refugee figures. The art piece is currently on display at the National Gallery in Prague until January 7, 2018.[150]

(2017) Permanent exhibit, unique setting of two Iron Trees from now on frame the Shrine of the Book in Jerusalem, Israel where Dead Sea Scrolls are preserved[151][152]

(2017) On the view in Israel Museum until the end of October 2017, Journey of Laziz is a video installation, showing mental breakdown and overall suffering of tiger living in the “world’s worst ZOO” in Gaza[151][152]

(2017) On view at the Park Avenue Armory through August 6, 2017, Hansel and Gretel is an installation exploring the theme of surveillance. The project, a collaboration of Ai Weiwei and architects Jacques Herzog and Pierre de Meuron, features surveillance cameras equipped with facial recognition software, near-infrared floor projections, tethered, autonomous drones and sonar beacons. A companion website includes a curatorial statement, artist biographies, a livestream of the installation and a timeline of surveillance technology from ancient to modern times.[153]

Originally posted here:

Ai Weiwei – Wikipedia

Posted in Ai

Wit.ai

Wit.ai makes it easy for developers to build applications and devices that you can talk or text to. Our vision is to empower developers with an open and extensible natural language platform. Wit.ai learns human language from every interaction, and leverages the community: what’s learned is shared across developers.

See the rest here:

Wit.ai

Posted in Ai

Artificial intelligence – Wikipedia

Artificial intelligence (AI, also machine intelligence, MI) is intelligence demonstrated by machines, in contrast to the natural intelligence (NI) displayed by humans and other animals. In computer science AI research is defined as the study of “intelligent agents”: any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.[1] Colloquially, the term “artificial intelligence” is applied when a machine mimics “cognitive” functions that humans associate with other human minds, such as “learning” and “problem solving”.

The scope of AI is disputed: as machines become increasingly capable, tasks considered as requiring “intelligence” are often removed from the definition, a phenomenon known as the AI effect, leading to the quip, “AI is whatever hasn’t been done yet.”[3] For instance, optical character recognition is frequently excluded from “artificial intelligence”, having become a routine technology.[4] Capabilities generally classified as AI as of 2017[update] include successfully understanding human speech, competing at the highest level in strategic game systems (such as chess and Go[6]), autonomous cars, intelligent routing in content delivery network and military simulations.

Artificial intelligence was founded as an academic discipline in 1956, and in the years since has experienced several waves of optimism,[7][8] followed by disappointment and the loss of funding (known as an “AI winter”),[9][10] followed by new approaches, success and renewed funding.[8][11] For most of its history, AI research has been divided into subfields that often fail to communicate with each other.[12] These sub-fields are based on technical considerations, such as particular goals (e.g. “robotics” or “machine learning”),[13] the use of particular tools (“logic” or “neural networks”), or deep philosophical differences.[14][15][16] Subfields have also been based on social factors (particular institutions or the work of particular researchers).[12]

The traditional problems (or goals) of AI research include reasoning, knowledge representation, planning, learning, natural language processing, perception and the ability to move and manipulate objects.[13] General intelligence is among the field’s long-term goals.[17] Approaches include statistical methods, computational intelligence, and traditional symbolic AI. Many tools are used in AI, including versions of search and mathematical optimization, neural networks and methods based on statistics, probability and economics. The AI field draws upon computer science, mathematics, psychology, linguistics, philosophy and many others.

The field was founded on the claim that human intelligence “can be so precisely described that a machine can be made to simulate it”.[18] This raises philosophical arguments about the nature of the mind and the ethics of creating artificial beings endowed with human-like intelligence, issues which have been explored by myth, fiction and philosophy since antiquity.[19] Some people also consider AI to be a danger to humanity if it progresses unabatedly.[20] Others believe that AI, unlike previous technological revolutions, will create a risk of mass unemployment.[21]

In the twenty-first century, AI techniques have experienced a resurgence following concurrent advances in computer power, large amounts of data, and theoretical understanding; and AI techniques have become an essential part of the technology industry, helping to solve many challenging problems in computer science.[22][11]

Thought-capable artificial beings appeared as storytelling devices in antiquity,[23] and have been common in fiction, as in Mary Shelley’s Frankenstein or Karel apek’s R.U.R. (Rossum’s Universal Robots).[24] These characters and their fates raised many of the same issues now discussed in the ethics of artificial intelligence.[19]

The study of mechanical or “formal” reasoning began with philosophers and mathematicians in antiquity. The study of mathematical logic led directly to Alan Turing’s theory of computation, which suggested that a machine, by shuffling symbols as simple as “0” and “1”, could simulate any conceivable act of mathematical deduction. This insight, that digital computers can simulate any process of formal reasoning, is known as the ChurchTuring thesis.[25] Along with concurrent discoveries in neurobiology, information theory and cybernetics, this led researchers to consider the possibility of building an electronic brain. Turing proposed that “if a human could not distinguish between responses from a machine and a human, the machine could be considered intelligent”. [26] The first work that is now generally recognized as AI was McCullouch and Pitts’ 1943 formal design for Turing-complete “artificial neurons”.

The field of AI research was born at a workshop at Dartmouth College in 1956.[28] Attendees Allen Newell (CMU), Herbert Simon (CMU), John McCarthy (MIT), Marvin Minsky (MIT) and Arthur Samuel (IBM) became the founders and leaders of AI research.[29] They and their students produced programs that the press described as “astonishing”: computers were learning checkers strategies (c. 1954)[31] (and by 1959 were reportedly playing better than the average human),[32] solving word problems in algebra, proving logical theorems (Logic Theorist, first run c. 1956) and speaking English.[33] By the middle of the 1960s, research in the U.S. was heavily funded by the Department of Defense[34] and laboratories had been established around the world.[35] AI’s founders were optimistic about the future: Herbert Simon predicted, “machines will be capable, within twenty years, of doing any work a man can do”. Marvin Minsky agreed, writing, “within a generation… the problem of creating ‘artificial intelligence’ will substantially be solved”.[7]

They failed to recognize the difficulty of some of the remaining tasks. Progress slowed and in 1974, in response to the criticism of Sir James Lighthill and ongoing pressure from the US Congress to fund more productive projects, both the U.S. and British governments cut off exploratory research in AI. The next few years would later be called an “AI winter”,[9] a period when obtaining funding for AI projects was difficult.

In the early 1980s, AI research was revived by the commercial success of expert systems,[37] a form of AI program that simulated the knowledge and analytical skills of human experts. By 1985 the market for AI had reached over a billion dollars. At the same time, Japan’s fifth generation computer project inspired the U.S and British governments to restore funding for academic research.[8] However, beginning with the collapse of the Lisp Machine market in 1987, AI once again fell into disrepute, and a second, longer-lasting hiatus began.[10]

In the late 1990s and early 21st century, AI began to be used for logistics, data mining, medical diagnosis and other areas.[22] The success was due to increasing computational power (see Moore’s law), greater emphasis on solving specific problems, new ties between AI and other fields (such as statistics, economics and mathematics), and a commitment by researchers to mathematical methods and scientific standards.[38] Deep Blue became the first computer chess-playing system to beat a reigning world chess champion, Garry Kasparov on 11 May 1997.

Advanced statistical techniques (loosely known as deep learning), access to large amounts of data and faster computers enabled advances in machine learning and perception. By the mid 2010s, machine learning applications were used throughout the world. In a Jeopardy! quiz show exhibition match, IBM’s question answering system, Watson, defeated the two greatest Jeopardy champions, Brad Rutter and Ken Jennings, by a significant margin. The Kinect, which provides a 3D bodymotion interface for the Xbox 360 and the Xbox One use algorithms that emerged from lengthy AI research[41] as do intelligent personal assistants in smartphones.[42] In March 2016, AlphaGo won 4 out of 5 games of Go in a match with Go champion Lee Sedol, becoming the first computer Go-playing system to beat a professional Go player without handicaps.[6][43] In the 2017 Future of Go Summit, AlphaGo won a three-game match with Ke Jie,[44] who at the time continuously held the world No. 1 ranking for two years.[45][46] This marked the completion of a significant milestone in the development of Artificial Intelligence as Go is an extremely complex game, more so than Chess.

According to Bloomberg’s Jack Clark, 2015 was a landmark year for artificial intelligence, with the number of software projects that use AI within Google increased from a “sporadic usage” in 2012 to more than 2,700 projects. Clark also presents factual data indicating that error rates in image processing tasks have fallen significantly since 2011.[47] He attributes this to an increase in affordable neural networks, due to a rise in cloud computing infrastructure and to an increase in research tools and datasets.[11] Other cited examples include Microsoft’s development of a Skype system that can automatically translate from one language to another and Facebook’s system that can describe images to blind people.[47]

A typical AI perceives its environment and takes actions that maximize its chance of successfully achieving its goals.[1] An AI’s intended goal function can be simple (“1 if the AI wins a game of Go, 0 otherwise”) or complex (“Do actions mathematically similar to the actions that got you rewards in the past”). Goals can be explicitly defined, or can be induced. If the AI is programmed for “reinforcement learning”, goals can be implicitly induced by rewarding some types of behavior and punishing others.[a] Alternatively, an evolutionary system can induce goals by using a “fitness function” to mutate and preferentially replicate high-scoring AI systems; this is similar to how animals evolved to innately desire certain goals such as finding food, or how dogs can be bred via artificial selection to possess desired traits. Some AI systems, such as nearest-neighbor, instead reason by analogy; these systems are not generally given goals, except to the degree that goals are somehow implicit in their training data. Such systems can still be benchmarked if the non-goal system is framed as a system whose “goal” is to successfully accomplish its narrow classification task.[50]

AI often revolves around the use of algorithms. An algorithm is a set of unambiguous instructions that a mechanical computer can execute.[b] A complex algorithm is often built on top of other, simpler, algorithms. A simple example of an algorithm is the following recipe for optimal play at tic-tac-toe:

Many AI algorithms are capable of learning from data; they can enhance themselves by learning new heuristics (strategies, or “rules of thumb”, that have worked well in the past), or can themselves write other algorithms. Some of the “learners” described below, including Bayesian networks, decision trees, and nearest-neighbor, could theoretically, if given infinite data, time, and memory, learn to approximate any function, including whatever combination of mathematical functions would best describe the entire world. These learners could therefore, in theory, derive all possible knowledge, by considering every possible hypothesis and matching it against the data. In practice, it is almost never possible to consider every possibility, because of the phenomenon of “combinatorial explosion”, where the amount of time needed to solve a problem grows exponentially. Much of AI research involves figuring out how to identify and avoid considering broad swaths of possibililities that are unlikely to be fruitful.[52] For example, when viewing a map and looking for the shortest driving route from Denver to New York in the East, one can in most cases skip looking at any path through San Francisco or other areas far to the West; thus, an AI wielding an pathfinding algorithm like A* can avoid the combinatorial explosion that would ensue if every possible route had to be ponderously considered in turn.[54]

The earliest (and easiest to understand) approach to AI was symbolism (such as formal logic): “If an otherwise healthy adult has a fever, then they may have influenza”. A second, more general, approach is Bayesian inference: “If the current patient has a fever, adjust the probability they have influenza in such-and-such way”. The third major approach, extremely popular in routine business AI applications, is analogizers such as SVM and nearest-neighbor: “After examining the records of known past patients whose temperature, symptoms, age, and other factors mostly match the current patient, X% of those patients turned out to have influenza”. A fourth approach is harder to intuitively understand, but is inspired by how the brain’s machinery works: the neural network approach uses artificial “neurons” that can learn by comparing itself to the desired output and altering the strengths of the connections between its internal neurons to “reinforce” connections that seemed to be useful. These four main approaches can overlap with each other and with evolutionary systems; for example, neural nets can learn to make inferences, to generalize, and to make analogies. Some systems implicitly or explicitly use multiple of these approaches, alongside many other AI and non-AI algorithms; the best approach is often different depending on the problem.[56]

Learning algorithms work on the basis that strategies, algorithms, and inferences that worked well in the past are likely to continue working well in the future. These inferences can be obvious, such as “since the sun rose every morning for the last 10,000 days, it will probably rise tomorrow morning as well”. They can be nuanced, such as “X% of families have geographically separate species with color variants, so there is an Y% chance that undiscovered black swans exist”. Learners also work on the basis of “Occam’s razor”: The simplest theory that explains the data is the likeliest. Therefore, to be successful, a learner must be designed such that it prefers simpler theories to complex theories, except in cases where the complex theory is proven substantially better. Settling on a bad, overly complex theory gerrymandered to fit all the past training data is known as overfitting. Many systems attempt to reduce overfitting by rewarding a theory in accordance with how well it fits the data, but penalizing the theory in accordance with how complex the theory is. Besides classic overfitting, learners can also disappoint by “learning the wrong lesson”. A toy example is that an image classifier trained only on pictures of brown horses and black cats might conclude that all brown patches are likely to be horses. A real-world example is that, unlike humans, current image classifiers don’t determine the spatial relationship between components of the picture; instead, they learn abstract patterns of pixels that humans are oblivious to, but that linearly correlate with images of certain types of real objects. Faintly superimposing such a pattern on a legitimate image results in an “adversarial” image that the system misclassifies.[c][59][60][61]

Compared with humans, existing AI lacks several features of human “commonsense reasoning”; most notably, humans have powerful mechanisms for reasoning about “nave physics” such as space, time, and physical interactions. This enables even young children to easily make inferences like “If I roll this pen off a table, it will fall on the floor”. Humans also have a powerful mechanism of “folk psychology” that helps them to interpret natural-language sentences such as “The city councilmen refused the demonstrators a permit because they advocated violence”. (A generic AI has difficulty inferring whether the councilmen or the demonstrators are the ones alleged to be advocating violence.)[64][65][66] This lack of “common knowledge” means that AI often makes different mistakes than humans make, in ways that can seem incomprehensible. For example, existing self-driving cars cannot reason about the location nor the intentions of pedestrians in the exact way that humans do, and instead must use non-human modes of reasoning to avoid accidents.[67][68][69]

The overall research goal of artificial intelligence is to create technology that allows computers and machines to function in an intelligent manner. The general problem of simulating (or creating) intelligence has been broken down into sub-problems. These consist of particular traits or capabilities that researchers expect an intelligent system to display. The traits described below have received the most attention.[13]

Early researchers developed algorithms that imitated step-by-step reasoning that humans use when they solve puzzles or make logical deductions.[70] By the late 1980s and 1990s, AI research had developed methods for dealing with uncertain or incomplete information, employing concepts from probability and economics.[71]

These algorithms proved to be insufficient for solving large reasoning problems, because they experienced a “combinatorial explosion”: they became exponentially slower as the problems grew larger.[52] In fact, even humans rarely use the step-by-step deduction that early AI research was able to model. They solve most of their problems using fast, intuitive judgements.[72] Modern statistical approaches to AI (e.g. neural networks) mimic this human ability to make a quick guess based on experience, solving many problems as people do. However, they are not capable of step-by-step deduction.

Knowledge representation[73] and knowledge engineering[74] are central to AI research. Many of the problems machines are expected to solve will require extensive knowledge about the world. Among the things that AI needs to represent are: objects, properties, categories and relations between objects;[75] situations, events, states and time;[76] causes and effects;[77] knowledge about knowledge (what we know about what other people know);[78] and many other, less well researched domains. A representation of “what exists” is an ontology: the set of objects, relations, concepts, and properties formally described so that software agents can interpret them. The semantics of these are captured as description logic concepts, roles, and individuals, and typically implemented as classes, properties, and individuals in the Web Ontology Language.[79] The most general ontologies are called upper ontologies, which attempt to provide a foundation for all other knowledge[80] by acting as mediators between domain ontologies that cover specific knowledge about a particular knowledge domain (field of interest or area of concern). Such formal knowledge representations are suitable for content-based indexing and retrieval, scene interpretation, clinical decision support, knowledge discovery via automated reasoning (inferring new statements based on explicitly stated knowledge), etc. Video events are often represented as SWRL rules, which can be used, among others, to automatically generate subtitles for constrained videos.[81]

Among the most difficult problems in knowledge representation are:

Intelligent agents must be able to set goals and achieve them.[88] They need a way to visualize the futurea representation of the state of the world and be able to make predictions about how their actions will change itand be able to make choices that maximize the utility (or “value”) of available choices.[89]

In classical planning problems, the agent can assume that it is the only system acting in the world, allowing the agent to be certain of the consequences of its actions.[90] However, if the agent is not the only actor, then it requires that the agent can reason under uncertainty. This calls for an agent that can not only assess its environment and make predictions, but also evaluate its predictions and adapt based on its assessment.[91]

Multi-agent planning uses the cooperation and competition of many agents to achieve a given goal. Emergent behavior such as this is used by evolutionary algorithms and swarm intelligence.[92]

Machine learning, a fundamental concept of AI research since the field’s inception,[93] is the study of computer algorithms that improve automatically through experience.[94][95]

Unsupervised learning is the ability to find patterns in a stream of input. Supervised learning includes both classification and numerical regression. Classification is used to determine what category something belongs in, after seeing a number of examples of things from several categories. Regression is the attempt to produce a function that describes the relationship between inputs and outputs and predicts how the outputs should change as the inputs change. In reinforcement learning[96] the agent is rewarded for good responses and punished for bad ones. The agent uses this sequence of rewards and punishments to form a strategy for operating in its problem space. These three types of learning can be analyzed in terms of decision theory, using concepts like utility. The mathematical analysis of machine learning algorithms and their performance is a branch of theoretical computer science known as computational learning theory.[citation needed]

Within developmental robotics, developmental learning approaches are elaborated upon to allow robots to accumulate repertoires of novel skills through autonomous self-exploration, social interaction with human teachers, and the use of guidance mechanisms (active learning, maturation, motor synergies, etc.).[97][98]

Natural language processing[101] gives machines the ability to read and understand human language. A sufficiently powerful natural language processing system would enable natural language user interfaces and the acquisition of knowledge directly from human-written sources, such as newswire texts. Some straightforward applications of natural language processing include information retrieval, text mining, question answering[102] and machine translation.[103]

A common method of processing and extracting meaning from natural language is through semantic indexing. Although these indexes require a large volume of user input, it is expected that increases in processor speeds and decreases in data storage costs will result in greater efficiency.

Machine perception[104] is the ability to use input from sensors (such as cameras, microphones, tactile sensors, sonar and others) to deduce aspects of the world. Computer vision[105] is the ability to analyze visual input. A few selected subproblems are speech recognition,[106] facial recognition and object recognition.[107]

The field of robotics[108] is closely related to AI. Intelligence is required for robots to handle tasks such as object manipulation[109] and navigation, with sub-problems such as localization, mapping, and motion planning. These systems require that an agent is able to: Be spatially cognizant of its surroundings, learn from and build a map of its environment, figure out how to get from one point in space to another, and execute that movement (which often involves compliant motion, a process where movement requires maintaining physical contact with an object).[111]

Affective computing is the study and development of systems that can recognize, interpret, process, and simulate human affects. It is an interdisciplinary field spanning computer sciences, psychology, and cognitive science. While the origins of the field may be traced as far back as the early philosophical inquiries into emotion, the more modern branch of computer science originated with Rosalind Picard’s 1995 paper on “affective computing”.[118][119] A motivation for the research is the ability to simulate empathy, where the machine would be able to interpret human emotions and adapts its behavior to give an appropriate response to those emotions.

Emotion and social skills[120] are important to an intelligent agent for two reasons. First, being able to predict the actions of others by understanding their motives and emotional states allow an agent to make better decisions. Concepts such as game theory, decision theory, necessitate that an agent be able to detect and model human emotions. Second, in an effort to facilitate humancomputer interaction, an intelligent machine may want to display emotions (even if it does not experience those emotions itself) to appear more sensitive to the emotional dynamics of human interaction.

A sub-field of AI addresses creativity both theoretically (the philosophical psychological perspective) and practically (the specific implementation of systems that generate novel and useful outputs).

Many researchers think that their work will eventually be incorporated into a machine with artificial general intelligence, combining all the skills mentioned above and even exceeding human ability in most or all these areas.[17][121] A few believe that anthropomorphic features like artificial consciousness or an artificial brain may be required for such a project.[122][123]

Many of the problems above may also require general intelligence, if machines are to solve the problems as well as people do. For example, even specific straightforward tasks, like machine translation, require that a machine read and write in both languages (NLP), follow the author’s argument (reason), know what is being talked about (knowledge), and faithfully reproduce the author’s original intent (social intelligence). A problem like machine translation is considered “AI-complete”, because all of these problems need to be solved simultaneously in order to reach human-level machine performance.

There is no established unifying theory or paradigm that guides AI research. Researchers disagree about many issues.[124] A few of the most long standing questions that have remained unanswered are these: should artificial intelligence simulate natural intelligence by studying psychology or neurobiology? Or is human biology as irrelevant to AI research as bird biology is to aeronautical engineering?[14] Can intelligent behavior be described using simple, elegant principles (such as logic or optimization)? Or does it necessarily require solving a large number of completely unrelated problems?[15] Can intelligence be reproduced using high-level symbols, similar to words and ideas? Or does it require “sub-symbolic” processing?[16] John Haugeland, who coined the term GOFAI (Good Old-Fashioned Artificial Intelligence), also proposed that AI should more properly be referred to as synthetic intelligence,[125] a term which has since been adopted by some non-GOFAI researchers.

Stuart Shapiro divides AI research into three approaches, which he calls computational psychology, computational philosophy, and computer science. Computational psychology is used to make computer programs that mimic human behavior.[128] Computational philosophy, is used to develop an adaptive, free-flowing computer mind.[128] Implementing computer science serves the goal of creating computers that can perform tasks that only people could previously accomplish.[128] Together, the humanesque behavior, mind, and actions make up artificial intelligence.

In the 1940s and 1950s, a number of researchers explored the connection between neurobiology, information theory, and cybernetics. Some of them built machines that used electronic networks to exhibit rudimentary intelligence, such as W. Grey Walter’s turtles and the Johns Hopkins Beast. Many of these researchers gathered for meetings of the Teleological Society at Princeton University and the Ratio Club in England.[129] By 1960, this approach was largely abandoned, although elements of it would be revived in the 1980s.

When access to digital computers became possible in the middle 1950s, AI research began to explore the possibility that human intelligence could be reduced to symbol manipulation. The research was centered in three institutions: Carnegie Mellon University, Stanford and MIT, and each one developed its own style of research. John Haugeland named these approaches to AI “good old fashioned AI” or “GOFAI”.[130] During the 1960s, symbolic approaches had achieved great success at simulating high-level thinking in small demonstration programs. Approaches based on cybernetics or neural networks were abandoned or pushed into the background.[131] Researchers in the 1960s and the 1970s were convinced that symbolic approaches would eventually succeed in creating a machine with artificial general intelligence and considered this the goal of their field.

Economist Herbert Simon and Allen Newell studied human problem-solving skills and attempted to formalize them, and their work laid the foundations of the field of artificial intelligence, as well as cognitive science, operations research and management science. Their research team used the results of psychological experiments to develop programs that simulated the techniques that people used to solve problems. This tradition, centered at Carnegie Mellon University would eventually culminate in the development of the Soar architecture in the middle 1980s.[132][133]

Unlike Newell and Simon, John McCarthy felt that machines did not need to simulate human thought, but should instead try to find the essence of abstract reasoning and problem solving, regardless of whether people used the same algorithms.[14] His laboratory at Stanford (SAIL) focused on using formal logic to solve a wide variety of problems, including knowledge representation, planning and learning.[134] Logic was also the focus of the work at the University of Edinburgh and elsewhere in Europe which led to the development of the programming language Prolog and the science of logic programming.[135]

Researchers at MIT (such as Marvin Minsky and Seymour Papert)[136] found that solving difficult problems in vision and natural language processing required ad-hoc solutions they argued that there was no simple and general principle (like logic) that would capture all the aspects of intelligent behavior. Roger Schank described their “anti-logic” approaches as “scruffy” (as opposed to the “neat” paradigms at CMU and Stanford).[15] Commonsense knowledge bases (such as Doug Lenat’s Cyc) are an example of “scruffy” AI, since they must be built by hand, one complicated concept at a time.[137]

When computers with large memories became available around 1970, researchers from all three traditions began to build knowledge into AI applications.[138] This “knowledge revolution” led to the development and deployment of expert systems (introduced by Edward Feigenbaum), the first truly successful form of AI software.[37] The knowledge revolution was also driven by the realization that enormous amounts of knowledge would be required by many simple AI applications.

By the 1980s progress in symbolic AI seemed to stall and many believed that symbolic systems would never be able to imitate all the processes of human cognition, especially perception, robotics, learning and pattern recognition. A number of researchers began to look into “sub-symbolic” approaches to specific AI problems.[16] Sub-symbolic methods manage to approach intelligence without specific representations of knowledge.

This includes embodied, situated, behavior-based, and nouvelle AI. Researchers from the related field of robotics, such as Rodney Brooks, rejected symbolic AI and focused on the basic engineering problems that would allow robots to move and survive.[139] Their work revived the non-symbolic viewpoint of the early cybernetics researchers of the 1950s and reintroduced the use of control theory in AI. This coincided with the development of the embodied mind thesis in the related field of cognitive science: the idea that aspects of the body (such as movement, perception and visualization) are required for higher intelligence.

Interest in neural networks and “connectionism” was revived by David Rumelhart and others in the middle of the 1980s.[140] Neural networks are an example of soft computing — they are solutions to problems which cannot be solved with complete logical certainty, and where an approximate solution is often sufficient. Other soft computing approaches to AI include fuzzy systems, evolutionary computation and many statistical tools. The application of soft computing to AI is studied collectively by the emerging discipline of computational intelligence.[141]

In the 1990s, AI researchers developed sophisticated mathematical tools to solve specific subproblems. These tools are truly scientific, in the sense that their results are both measurable and verifiable, and they have been responsible for many of AI’s recent successes. The shared mathematical language has also permitted a high level of collaboration with more established fields (like mathematics, economics or operations research). Stuart Russell and Peter Norvig describe this movement as nothing less than a “revolution” and “the victory of the neats”.[38] Critics argue that these techniques (with few exceptions) are too focused on particular problems and have failed to address the long-term goal of general intelligence. There is an ongoing debate about the relevance and validity of statistical approaches in AI, exemplified in part by exchanges between Peter Norvig and Noam Chomsky.

In the course of 60 or so years of research, AI has developed a large number of tools to solve the most difficult problems in computer science. A few of the most general of these methods are discussed below.

Many problems in AI can be solved in theory by intelligently searching through many possible solutions:[149] Reasoning can be reduced to performing a search. For example, logical proof can be viewed as searching for a path that leads from premises to conclusions, where each step is the application of an inference rule.[150] Planning algorithms search through trees of goals and subgoals, attempting to find a path to a target goal, a process called means-ends analysis.[151] Robotics algorithms for moving limbs and grasping objects use local searches in configuration space.[109] Many learning algorithms use search algorithms based on optimization.

Simple exhaustive searches[152] are rarely sufficient for most real world problems: the search space (the number of places to search) quickly grows to astronomical numbers. The result is a search that is too slow or never completes. The solution, for many problems, is to use “heuristics” or “rules of thumb” that prioritize choices in favor of those that are more likely to reach a goal, and to do so in a shorter number of steps. In some search methodologies heuristics can also serve to entirely eliminate some choices that are unlikely to lead to a goal (called “pruning the search tree”). Heuristics supply the program with a “best guess” for the path on which the solution lies.[153] Heuristics limit the search for solutions into a smaller sample size.

A very different kind of search came to prominence in the 1990s, based on the mathematical theory of optimization. For many problems, it is possible to begin the search with some form of a guess and then refine the guess incrementally until no more refinements can be made. These algorithms can be visualized as blind hill climbing: we begin the search at a random point on the landscape, and then, by jumps or steps, we keep moving our guess uphill, until we reach the top. Other optimization algorithms are simulated annealing, beam search and random optimization.[154]

Evolutionary computation uses a form of optimization search. For example, they may begin with a population of organisms (the guesses) and then allow them to mutate and recombine, selecting only the fittest to survive each generation (refining the guesses). Forms of evolutionary computation include swarm intelligence algorithms (such as ant colony or particle swarm optimization)[155] and evolutionary algorithms (such as genetic algorithms, gene expression programming, and genetic programming).[156]

Logic[157] is used for knowledge representation and problem solving, but it can be applied to other problems as well. For example, the satplan algorithm uses logic for planning[158] and inductive logic programming is a method for learning.[159]

Several different forms of logic are used in AI research. Propositional or sentential logic[160] is the logic of statements which can be true or false. First-order logic[161] also allows the use of quantifiers and predicates, and can express facts about objects, their properties, and their relations with each other. Fuzzy logic,[162] is a version of first-order logic which allows the truth of a statement to be represented as a value between 0 and 1, rather than simply True (1) or False (0). Fuzzy systems can be used for uncertain reasoning and have been widely used in modern industrial and consumer product control systems. Subjective logic[citation needed] models uncertainty in a different and more explicit manner than fuzzy-logic: a given binomial opinion satisfies belief + disbelief + uncertainty = 1 within a Beta distribution. By this method, ignorance can be distinguished from probabilistic statements that an agent makes with high confidence.

Default logics, non-monotonic logics and circumscription[83] are forms of logic designed to help with default reasoning and the qualification problem. Several extensions of logic have been designed to handle specific domains of knowledge, such as: description logics;[75] situation calculus, event calculus and fluent calculus (for representing events and time);[76] causal calculus;[77] belief calculus;[163] and modal logics.[78]

Many problems in AI (in reasoning, planning, learning, perception and robotics) require the agent to operate with incomplete or uncertain information. AI researchers have devised a number of powerful tools to solve these problems using methods from probability theory and economics.[164]

Bayesian network[165] is a very general tool that can be used for a large number of problems: reasoning (using the Bayesian inference algorithm),[166] learning (using the expectation-maximization algorithm),[d][168] planning (using decision networks)[169] and perception (using dynamic Bayesian networks).[170] Bayesian networks are used in AdSense to choose what ads to place and on XBox Live to rate and match players. Probabilistic algorithms can also be used for filtering, predicting, smoothing and finding explanations for streams of data, helping perception systems to analyze processes that occur over time (e.g., hidden Markov models or Kalman filters).[170]

A key concept from the science of economics is “utility”: a measure of how valuable something is to an intelligent agent. Precise mathematical tools have been developed that analyze how an agent can make choices and plan, using decision theory, decision analysis,[172] and information value theory.[89] These tools include models such as Markov decision processes,[173] dynamic decision networks,[170] game theory and mechanism design.[174]

The simplest AI applications can be divided into two types: classifiers (“if shiny then diamond”) and controllers (“if shiny then pick up”). Controllers do, however, also classify conditions before inferring actions, and therefore classification forms a central part of many AI systems. Classifiers are functions that use pattern matching to determine a closest match. They can be tuned according to examples, making them very attractive for use in AI. These examples are known as observations or patterns. In supervised learning, each pattern belongs to a certain predefined class. A class can be seen as a decision that has to be made. All the observations combined with their class labels are known as a data set. When a new observation is received, that observation is classified based on previous experience.[175]

A classifier can be trained in various ways; there are many statistical and machine learning approaches. The decision tree[176] is perhaps the most widely used machine learning algorithm. Other widely used classifiers are the neural network,[178] k-nearest neighbor algorithm,[e][180] kernel methods such as the support vector machine (SVM),[f][182] Gaussian mixture model[183] and the extremely popular naive Bayes classifier.[g][185] The performance of these classifiers have been compared over a wide range of tasks. Classifier performance depends greatly on the characteristics of the data to be classified. There is no single classifier that works best on all given problems; this is also referred to as the “no free lunch” theorem. Determining a suitable classifier for a given problem is still more an art than science.[186]

Neural networks, or neural nets, were inspired by the architecture of neurons in the human brain. A simple “neuron” N accepts input from multiple other neurons, each of which, when activated (or “fired”), cast a weighted “vote” for or against whether neuron N should itself activate. Learning requires an algorithm to adjust these weights based on the training data; one simple algorithm (dubbed “fire together, wire together”) is to increase the weight between two connected neurons when the activation of one triggers the successful activation of another. The net forms “concepts” that are distributed among a subnetwork of shared[h] neurons that tend to fire together; a concept meaning “leg” might be coupled with a subnetwork meaning “foot” that includes the sound for “foot”. Neurons have a continuous spectrum of activation; in addition, neurons can process inputs in a nonlinear way rather than weighing straightforward votes. Modern neural nets can learn both continuous functions and, surprisingly, digital logical operations. Neural networks’ early successes included predicting the stock market and (in 1995) a mostly self-driving car.[i] In the 2010s, advances in neural networks using deep learning thrust AI into widespread public consciousness and contributed to an enormous upshift in corporate AI spending; for example, AI-related M&A in 2017 was over 25 times as large as in 2015.[188][189]

The study of non-learning artificial neural networks[178] began in the decade before the field of AI research was founded, in the work of Walter Pitts and Warren McCullouch. Frank Rosenblatt invented the perceptron, a learning network with a single layer, similar to the old concept of linear regression. Early pioneers also include Alexey Grigorevich Ivakhnenko, Teuvo Kohonen, Stephen Grossberg, Kunihiko Fukushima, Christoph von der Malsburg, David Willshaw, Shun-Ichi Amari, Bernard Widrow, John Hopfield, Eduardo R. Caianiello, and others.

The main categories of networks are acyclic or feedforward neural networks (where the signal passes in only one direction) and recurrent neural networks (which allow feedback and short-term memories of previous input events). Among the most popular feedforward networks are perceptrons, multi-layer perceptrons and radial basis networks.[190] Neural networks can be applied to the problem of intelligent control (for robotics) or learning, using such techniques as Hebbian learning (“fire together, wire together”), GMDH or competitive learning.[191]

Today, neural networks are often trained by the backpropagation algorithm, which had been around since 1970 as the reverse mode of automatic differentiation published by Seppo Linnainmaa,[192][193] and was introduced to neural networks by Paul Werbos.[194][195][196]

Hierarchical temporal memory is an approach that models some of the structural and algorithmic properties of the neocortex.[197]

In short, most neural networks use some form of gradient descent on a hand-created neural topology. However, some research groups, such as Uber, argue that simple neuroevolution to mutate new neural network topologies and weights may be competitive with sophisticated gradient descent approaches. One advantage of neuroevolution is that it may be less prone to get caught in “dead ends”.[198]

Deep learning is any artificial neural network that can learn a long chain of causal links. For example, a feedforward network with six hidden layers can learn a seven-link causal chain (six hidden layers + output layer) and has a “credit assignment path” (CAP) depth of seven. Many deep learning systems need to be able to learn chains ten or more causal links in length.[199] Deep learning has transformed many important subfields of artificial intelligence, including computer vision, speech recognition, natural language processing and others.[200][201][199]

According to one overview,[202] the expression “Deep Learning” was introduced to the Machine Learning community by Rina Dechter in 1986[203] and gained traction after Igor Aizenberg and colleagues introduced it to Artificial Neural Networks in 2000.[204] The first functional Deep Learning networks were published by Alexey Grigorevich Ivakhnenko and V. G. Lapa in 1965.[205][pageneeded] These networks are trained one layer at a time. Ivakhnenko’s 1971 paper[206] describes the learning of a deep feedforward multilayer perceptron with eight layers, already much deeper than many later networks. In 2006, a publication by Geoffrey Hinton and Ruslan Salakhutdinov introduced another way of pre-training many-layered feedforward neural networks (FNNs) one layer at a time, treating each layer in turn as an unsupervised restricted Boltzmann machine, then using supervised backpropagation for fine-tuning. Similar to shallow artificial neural networks, deep neural networks can model complex non-linear relationships. Over the last few years, advances in both machine learning algorithms and computer hardware have led to more efficient methods for training deep neural networks that contain many layers of non-linear hidden units and a very large output layer.[208]

Deep learning often uses convolutional neural networks (CNNs), whose origins can be traced back to the Neocognitron introduced by Kunihiko Fukushima in 1980.[209] In 1989, Yann LeCun and colleagues applied backpropagation to such an architecture. In the early 2000s, in an industrial application CNNs already processed an estimated 10% to 20% of all the checks written in the US.[210] Since 2011, fast implementations of CNNs on GPUs have won many visual pattern recognition competitions.[199]

CNNs with 12 convolutional layers were used in conjunction with reinforcement learning by Deepmind’s “AlphaGo Lee”, the program that beat a top Go champion in 2016.[211]

Early on, deep learning was also applied to sequence learning with recurrent neural networks (RNNs)[212] which are in theory Turing complete[213] and can run arbitrary programs to process arbitrary sequences of inputs. The depth of an RNN is unlimited and depends on the length of its input sequence; thus, an RNN is an example of deep learning.[199] RNNs can be trained by gradient descent[214][215][216] but suffer from the vanishing gradient problem.[200][217] In 1992, it was shown that unsupervised pre-training of a stack of recurrent neural networks can speed up subsequent supervised learning of deep sequential problems.[218]

Numerous researchers now use variants of a deep learning recurrent NN called the long short-term memory (LSTM) network published by Hochreiter & Schmidhuber in 1997.[219] LSTM is often trained by Connectionist Temporal Classification (CTC).[220] At Google, Microsoft and Baidu this approach has revolutionised speech recognition.[221][222][223] For example, in 2015, Google’s speech recognition experienced a dramatic performance jump of 49% through CTC-trained LSTM, which is now available through Google Voice to billions of smartphone users.[224] Google also used LSTM to improve machine translation,[225] Language Modeling[226] and Multilingual Language Processing.[227] LSTM combined with CNNs also improved automatic image captioning[228] and a plethora of other applications.

Early symbolic AI inspired Lisp[229] and Prolog,[230] which dominated early AI programming. Modern AI development often uses mainstream languages such as Python or C++,[231] or niche languages such as Wolfram Language.[232]

In 1950, Alan Turing proposed a general procedure to test the intelligence of an agent now known as the Turing test. This procedure allows almost all the major problems of artificial intelligence to be tested. However, it is a very difficult challenge and at present all agents fail.[233]

Artificial intelligence can also be evaluated on specific problems such as small problems in chemistry, hand-writing recognition and game-playing. Such tests have been termed subject matter expert Turing tests. Smaller problems provide more achievable goals and there are an ever-increasing number of positive results.[citation needed]

For example, performance at draughts (i.e. checkers) is optimal,[citation needed] performance at chess is high-human and nearing super-human (see computer chess:computers versus human) and performance at many everyday tasks (such as recognizing a face or crossing a room without bumping into something) is sub-human.

A quite different approach measures machine intelligence through tests which are developed from mathematical definitions of intelligence. Examples of these kinds of tests start in the late nineties devising intelligence tests using notions from Kolmogorov complexity and data compression.[234] Two major advantages of mathematical definitions are their applicability to nonhuman intelligences and their absence of a requirement for human testers.

A derivative of the Turing test is the Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA). As the name implies, this helps to determine that a user is an actual person and not a computer posing as a human. In contrast to the standard Turing test, CAPTCHA is administered by a machine and targeted to a human as opposed to being administered by a human and targeted to a machine. A computer asks a user to complete a simple test then generates a grade for that test. Computers are unable to solve the problem, so correct solutions are deemed to be the result of a person taking the test. A common type of CAPTCHA is the test that requires the typing of distorted letters, numbers or symbols that appear in an image undecipherable by a computer.

AI is relevant to any intellectual task. Modern artificial intelligence techniques are pervasive and are too numerous to list here. Frequently, when a technique reaches mainstream use, it is no longer considered artificial intelligence; this phenomenon is described as the AI effect.

High-profile examples of AI include autonomous vehicles (such as drones and self-driving cars), medical diagnosis, creating art (such as poetry), proving mathematical theorems, playing games (such as Chess or Go), search engines (such as Google search), online assistants (such as Siri), image recognition in photographs, spam filtering, prediction of judicial decisions[238] and targeting online advertisements.[239][240]

With social media sites overtaking TV as a source for news for young people and news organisations increasingly reliant on social media platforms for generating distribution,[241] major publishers now use artificial intelligence (AI) technology to post stories more effectively and generate higher volumes of traffic.[242]

There are a number of competitions and prizes to promote research in artificial intelligence. The main areas promoted are: general machine intelligence, conversational behavior, data-mining, robotic cars, robot soccer and games.

Artificial intelligence is breaking into the healthcare industry by assisting doctors. According to Bloomberg Technology, Microsoft has developed AI to help doctors find the right treatments for cancer.[243] There is a great amount of research and drugs developed relating to cancer. In detail, there are more than 800 medicines and vaccines to treat cancer. This negatively affects the doctors, because there are too many options to choose from, making it more difficult to choose the right drugs for the patients. Microsoft is working on a project to develop a machine called “Hanover”. Its goal is to memorize all the papers necessary to cancer and help predict which combinations of drugs will be most effective for each patient. One project that is being worked on at the moment is fighting myeloid leukemia, a fatal cancer where the treatment has not improved in decades. Another study was reported to have found that artificial intelligence was as good as trained doctors in identifying skin cancers.[244] Another study is using artificial intelligence to try and monitor multiple high-risk patients, and this is done by asking each patient numerous questions based on data acquired from live doctor to patient interactions.[245]

According to CNN, there was a recent study by surgeons at the Children’s National Medical Center in Washington which successfully demonstrated surgery with an autonomous robot. The team supervised the robot while it performed soft-tissue surgery, stitching together a pig’s bowel during open surgery, and doing so better than a human surgeon, the team claimed.[246] IBM has created its own artificial intelligence computer, the IBM Watson, which has beaten human intelligence (at some levels). Watson not only won at the game show Jeopardy! against former champions,[247] but, was declared a hero after successfully diagnosing a women who was suffering from leukemia.[248]

Advancements in AI have contributed to the growth of the automotive industry through the creation and evolution of self-driving vehicles. As of 2016, there are over 30 companies utilizing AI into the creation of driverless cars. A few companies involved with AI include Tesla, Google, and Apple.[249]

Many components contribute to the functioning of self-driving cars. These vehicles incorporate systems such as braking, lane changing, collision prevention, navigation and mapping. Together, these systems, as well as high performance computers, are integrated into one complex vehicle.[250]

Recent developments in autonomous automobiles have made the innovation of self-driving trucks possible, though they are still in the testing phase. The UK government has passed legislation to begin testing of self-driving truck platoons in 2018.[251] Self-driving truck platoons are a fleet of self-driving trucks following the lead of one non-self-driving truck, so the truck platoons aren’t entirely autonomous yet. Meanwhile, the Daimler, a German automobile corporation, is testing the Freightliner Inspiration which is a semi-autonomous truck that will only be used on the highway.[252]

One main factor that influences the ability for a driver-less automobile to function is mapping. In general, the vehicle would be pre-programmed with a map of the area being driven. This map would include data on the approximations of street light and curb heights in order for the vehicle to be aware of its surroundings. However, Google has been working on an algorithm with the purpose of eliminating the need for pre-programmed maps and instead, creating a device that would be able to adjust to a variety of new surroundings.[253] Some self-driving cars are not equipped with steering wheels or brake pedals, so there has also been research focused on creating an algorithm that is capable of maintaining a safe environment for the passengers in the vehicle through awareness of speed and driving conditions.[254]

Another factor that is influencing the ability for a driver-less automobile is the safety of the passenger. To make a driver-less automobile, engineers must program it to handle high risk situations. These situations could include a head on collision with pedestrians. The car’s main goal should be to make a decision that would avoid hitting the pedestrians and saving the passengers in the car. But there is a possibility the car would need to make a decision that would put someone in danger. In other words, the car would need to decide to save the pedestrians or the passengers.[255] The programing of the car in these situations is crucial to a successful driver-less automobile.

Financial institutions have long used artificial neural network systems to detect charges or claims outside of the norm, flagging these for human investigation. The use of AI in banking can be traced back to 1987 when Security Pacific National Bank in US set-up a Fraud Prevention Task force to counter the unauthorised use of debit cards. Programs like Kasisto and Moneystream are using AI in financial services.

Banks use artificial intelligence systems today to organize operations, maintain book-keeping, invest in stocks, and manage properties. AI can react to changes overnight or when business is not taking place.[256] In August 2001, robots beat humans in a simulated financial trading competition.[257] AI has also reduced fraud and financial crimes by monitoring behavioral patterns of users for any abnormal changes or anomalies.[258]

The use of AI machines in the market in applications such as online trading and decision making has changed major economic theories.[259] For example, AI based buying and selling platforms have changed the law of supply and demand in that it is now possible to easily estimate individualized demand and supply curves and thus individualized pricing. Furthermore, AI machines reduce information asymmetry in the market and thus making markets more efficient while reducing the volume of trades. Furthermore, AI in the markets limits the consequences of behavior in the markets again making markets more efficient. Other theories where AI has had impact include in rational choice, rational expectations, game theory, Lewis turning point, portfolio optimization and counterfactual thinking.

In video games, artificial intelligence is routinely used to generate dynamic purposeful behavior in non-player characters (NPCs). In addition, well-understood AI techniques are routinely used for pathfinding. Some researchers consider NPC AI in games to be a “solved problem” for most production tasks. Games with more atypical AI include the AI director of Left 4 Dead (2008) and the neuroevolutionary training of platoons in Supreme Commander 2 (2010).[260][261]

Worldwide annual military spending on robotics rose from 5.1 billion USD in 2010 to 7.5 billion USD in 2015.[262][263] Military drones capable of autonomous action are widely considered a useful asset. In 2017, Vladimir Putin stated that “Whoever becomes the leader in (artificial intelligence) will become the ruler of the world”.[264][265] Many artificial intelligence researchers seek to distance themselves from military applications of AI.[266]

A platform (or “computing platform”) is defined as “some sort of hardware architecture or software framework (including application frameworks), that allows software to run”. As Rodney Brooks pointed out many years ago, it is not just the artificial intelligence software that defines the AI features of the platform, but rather the actual platform itself that affects the AI that results, i.e., there needs to be work in AI problems on real-world platforms rather than in isolation.

See more here:

Artificial intelligence – Wikipedia

Posted in Ai

AI File – What is it and how do I open it?

Did your computer fail to open an AI file? We explain what AI files are and recommend software that we know can open or convert your AI files.

AI is the acronym for Adobe Illustrator. Files that have the .ai extension are drawing files that the Adobe Illustrator application has created.

The Adobe Illustrator application was developed by Adobe Systems. The files created by this application are composed of paths that are connected by points and are saved in vector format. The technology used to create these files allows the user to re-size the AI image without losing any of the image’s quality.

Some third-party programs allow users to “rastersize” the images created in Adobe Illustrator, which allows them to convert the AI file into bitmap format. While this may make the file size smaller and easier to open across multiple applications, some of the file quality may be lost in the process.

See original here:

AI File – What is it and how do I open it?

Posted in Ai

AI Platform | Microsoft Azure

Sign up for the AI newsletter to follow the latest Microsoft AI news, features, events, and community activities surrounding Cognitive Services, Bot Framework, Machine Learning, and Cognitive Toolkit.

First name

Last name

Email

Country/region Afghanistan land Islands Albania Algeria American Samoa Andorra Angola Anguilla Antarctica Antigua and Barbuda Argentina Armenia Aruba Australia Austria Azerbaijan Bahamas Bahrain Bailiwick of Guernsey Bailiwick of Jersey Bangladesh Barbados Belarus Belgium Belize Benin Bermuda Bhutan Bolivia Bonaire Bosnia and Herzegovina Botswana Bouvet Island Brazil British Indian Ocean Territory British Virgin Islands Brunei Darussalam Bulgaria Burkina Faso Cameroon Canada Cayman Islands Central African Republic Chile China Colombia Commonwealth of Dominica Commonwealth of the Northern Mariana Islands Cook Islands Cooperative Republic of Guyana Costa Rica Cte d’Ivoire Country of Sint Maarten Croatia Curaao Cyprus Czech Republic Democratic People’s Republic of Korea Democratic Republic of So Tom and Prncipe Democratic Republic of the Congo Democratic Republic of Timor-Leste Denmark Department of Guadeloupe Department of Guiana Department of Martinique Department of Mayotte Department of Runion Dominican Republic Ecuador Egypt El Salvador Estonia Ethiopia Falkland Islands Faroe Islands Federal Republic of Somalia Federated States of Micronesia Fiji Finland France Gabonese Republic Georgia Germany Ghana Gibraltar Greece Greenland Grenada Guatemala Honduras Hong Kong Hong Kong Hungary Iceland Independent State of Papua New Guinea Independent State of Samoa India Indonesia Iraq Ireland Islamic Republic of Iran Islamic Republic of Mauritania Isle of Man Israel Italy Jamaica Jan Mayen Japan Jordan Kazakhstan Kenya Kingdom of Cambodia Kingdom of Lesotho Kingdom of Swaziland Kingdom of Tonga Korea Kosovo Kuwait Kyrgyzstan Lao People’s Democratic Republic Latvia Lebanon Libya Liechtenstein Lithuania Luxembourg Macao Macedonia (FYRO) Malaysia Malta Mauritius Mexico Moldova Monaco Mongolia Montenegro Montserrat Morocco Namibia Nepal Netherlands New Zealand Nicaragua Nigeria Niue Norway Oman Overseas Collectivity of Saint Barthlemy Overseas Collectivity of Saint Martin Overseas Collectivity of Saint Pierre and Miquelon Overseas Collectivity of the Wallis and Futuna Islands Overseas Lands of French Polynesia Pakistan Palestinian Authority Panama Paraguay Peru Philippines Pitcairn, Henderson, Ducie, and Oeno Islands Poland Portugal Public Body of Saba Public Body of Sint Eustatius Puerto Rico Qatar Republic of Burundi Republic of Cabo Verde Republic of Chad Republic of Cuba Republic of Djibouti Republic of Equatorial Guinea Republic of Guinea Republic of Guinea-Bissau Republic of Haiti Republic of Kiribati Republic of Liberia Republic of Madagascar Republic of Malawi Republic of Maldives Republic of Mali Republic of Mozambique Republic of Nauru Republic of Niger Republic of Palau Republic of San Marino Republic of Seychelles Republic of Sierra Leone Republic of South Sudan Republic of Suriname Republic of the Congo Republic of The Gambia Republic of the Marshall Islands Republic of the Sudan Republic of the Union of Myanmar Republic of Vanuatu Romania Russia Rwanda Saint Helena, Ascension, and Tristan da Cunha Saint Kitts and Nevis Saint Lucia Saint Vincent and the Grenadines Saudi Arabia Senegal Serbia Singapore Slovakia Slovenia Solomon Islands South Africa South Georgia and the South Sandwich Islands Spain Sri Lanka State of Eritrea State of the Vatican City Svalbard Sweden Switzerland Syrian Arab Republic Taiwan Tajikistan Tanzania Territory of Christmas Island Territory of Cocos (Keeling) Islands Territory of Guam Territory of Heard Island and McDonald Islands Territory of New Caledonia and Dependencies Territory of Norfolk Island Territory of the French Southern and Antarctic Lands Thailand Togolese Republic Tokelau Trinidad and Tobago Tunisia Turkey Turkmenistan Turks and Caicos Islands Tuvalu Uganda Ukraine Union of the Comoros United Arab Emirates United Kingdom United States United States Minor Outlying Islands Uruguay Uzbekistan Venezuela Vietnam Virgin Islands of the United States Yemen Zambia Zimbabwe

Microsoft may use your contact information to provide updates and special offers about Microsoft Azure. You can unsubscribe at any time. To learn more, read the privacy statement.

I accept

Subscribe

Go here to read the rest:

AI Platform | Microsoft Azure

Posted in Ai

Ai Weiwei – Wikipedia

Ai Weiwei (Chinese: ; pinyin: i Wiwi, English pronunciation(helpinfo); born 28 August 1957 in Beijing) is a Chinese contemporary artist and activist. His father’s (Ai Qing) original surname was written Jiang ().[1][2][3] Ai collaborated with Swiss architects Herzog & de Meuron as the artistic consultant on the Beijing National Stadium for the 2008 Olympics.[4] As a political activist, he has been highly and openly critical of the Chinese Government’s stance on democracy and human rights. He has investigated government corruption and cover-ups, in particular the Sichuan schools corruption scandal following the collapse of so-called “tofu-dreg schools” in the 2008 Sichuan earthquake.[5] In 2011, following his arrest at Beijing Capital International Airport on 3 April, he was held for 81 days without any official charges being filed; officials alluded to their allegations of “economic crimes”.[6]

Ai’s father was the Chinese poet Ai Qing,[7] who was denounced during the Anti-Rightist Movement. In 1958, the family was sent to a labour camp in Beidahuang, Heilongjiang, when Ai was one year old. They were subsequently exiled to Shihezi, Xinjiang in 1961, where they lived for 16 years. Upon Mao Zedong’s death and the end of the Cultural Revolution, the family returned to Beijing in 1976.[8]

In 1978, Ai enrolled in the Beijing Film Academy and studied animation.[9] In 1978, he was one of the founders of the early avant garde art group the “Stars”, together with Ma Desheng, Wang Keping, Huang Rui, Li Shuang, Ah Cheng and Qu Leilei. The group disbanded in 1983,[10] yet Ai participated in regular Stars group shows, The Stars: Ten Years, 1989 (Hanart Gallery, Hong Kong and Taipei), and a retrospective exhibition in Beijing in 2007: Origin Point (Today Art Museum, Beijing).[citation needed]

From 1981 to 1993, he lived in the United States. For the first few years, Ai lived in Philadelphia and San Francisco. He studied English at the University of Pennsylvania and the University of California, Berkeley.[12] Later, he moved to New York City.[10] He studied briefly at Parsons School of Design.[13] Ai attended the Art Students League of New York from 1983 to 1986, where he studied with Bruce Dorfman, Knox Martin and Richard Pousette-Dart.[14] He later dropped out of school, and made a living out of drawing street portraits and working odd jobs. During this period, he gained exposure to the works of Marcel Duchamp, Andy Warhol, and Jasper Johns, and began creating conceptual art by altering readymade objects.

Ai befriended beat poet Allen Ginsberg while living in New York, following a chance meeting at a poetry reading where Ginsberg read out several poems about China. Ginsberg had travelled to China and met with Ai’s father, the noted poet Ai Qing, and consequently Ginsberg and Ai became friends.[15]

When he was living in the East Village (from 1983 to 1993), Ai carried a camera with him all the time and would take pictures of his surroundings wherever he was. The resulting collection of photos were later selected and is now known as the New York Photographs.[16]

At the same time, Ai became fascinated by blackjack card games and frequented Atlantic City casinos. He is still regarded in gambling circles as a top tier professional blackjack player according to an article published on blackjackchamp.com.[17][18][19]

In 1993, Ai returned to China after his father became ill.[20] He helped establish the experimental artists’ Beijing East Village and co-published a series of three books about this new generation of artists with Chinese curator Feng Boyi: Black Cover Book (1994), White Cover Book (1995), and Gray Cover Book (1997).[21]

In 1999, Ai moved to Caochangdi, in the northeast of Beijing, and built a studio house his first architectural project. Due to his interest in architecture, he founded the architecture studio FAKE Design, in 2003.[22] In 2000, he co-curated the art exhibition Fuck Off with curator Feng Boyi in Shanghai, China.[23]

Ai is married to artist Lu Qing,[24] and has a son from an extramarital relationship.[25]

In 2005, Ai was invited to start blogging by Sina Weibo, the biggest internet platform in China. He posted his first blog on 19 November. For four years, he “turned out a steady stream of scathing social commentary, criticism of government policy, thoughts on art and architecture, and autobiographical writings.”[26] The blog was shut down by Sina on 28 May 2009. Ai then turned to Twitter and wrote prolifically on the platform, claiming at least eight hours online every day. He wrote almost exclusively in Chinese using the account @aiww.[citation needed] As of 31 December 2013, Ai has declared that he would stop tweeting but the account remains active in forms of retweets and Instagram posts.[27]

Ai supported the Amnesty International petition for Iranian filmmaker Hossein Rajabian and his brother, musician Mehdi Rajabian, and released the news on his Twitter pages.[28]

Ten days after the 8.0-magnitude earthquake in Sichuan province on 12 May 2008, Ai led a team to survey and film the post-quake conditions in various disaster zones. In response to the government’s lack of transparency in revealing names of students who perished in the earthquake due to substandard school campus constructions, Ai recruited volunteers online and launched a “Citizens’ Investigation” to compile names and information of the student victims. On 20 March 2009, he posted a blog titled “Citizens’ Investigation” and wrote: “To remember the departed, to show concern for life, to take responsibility, and for the potential happiness of the survivors, we are initiating a “Citizens’ Investigation.” We will seek out the names of each departed child, and we will remember them.”[29]

As of 14 April 2009, the list had accumulated 5,385 names.[30] Ai published the collected names as well as numerous articles documenting the investigation on his blog which was shut down by Chinese authorities in May 2009.[31] He also posted his list of names of schoolchildren who died on the wall of his office at FAKE Design in Beijing.[32]

Ai suffered headaches and claimed he had difficulty concentrating on his work since returning from Chengdu in August 2009, where he was beaten by the police for trying to testify for Tan Zuoren, a fellow investigator of the shoddy construction and student casualties in the earthquake. On 14 September 2009, Ai was diagnosed to be suffering internal bleeding in a hospital in Munich, Germany, and the doctor arranged for emergency brain surgery.[33] The cerebral hemorrhage is believed to be linked to the police attack.[34][35]

According to the Financial Times, in an attempt to force Ai to leave the country, two accounts used by him had been hacked in a sophisticated attack on Google in China dubbed Operation Aurora, their contents read and copied; his bank accounts were investigated by state security agents who claimed he was under investigation for “unspecified suspected crimes”.[36]

In November 2010, Ai was placed under house arrest by the Chinese police. He said this was to prevent the planned party marking the demolition of his newly built Shanghai studio.[37]

The building was designed and built by Ai upon encouragement and persuasion from a “high official [from Shanghai]” as part of a new cultural area designated by Shanghai Municipal authorities; Ai would have used it as a studio and to teach architecture courses. But now Ai has been accused of erecting the structure without the necessary planning permission and a demolition notice has been ordered, even though, Ai said, officials had been extremely enthusiastic, and the entire application and planning process was “under government supervision”. According to Ai, a number of artists were invited to build new studios in this area of Shanghai because officials wanted to create a cultural area.[38]

On 3 November 2010, Ai said the government had informed him two months earlier that the newly completed studio would be knocked down because it was illegal. Ai complained that this was unfair, as he was “the only one singled out to have my studio destroyed”. The Guardian reported Ai saying Shanghai municipal authorities were “frustrated” by documentaries on subjects they considered sensitive:[38] two of the better known ones featured Shanghai resident Feng Zhenghu, who lived in forced exile for three months in Narita Airport, Tokyo; another well-known documentary focused on Yang Jia, who murdered six Shanghai police officers.[39]

In the end, the party took place without Ai’s presence; his supporters feasted on river crab, an allusion to “harmony”, and a euphemism used to jeer official censorship. Ai was released from house arrest the next day.[40]

Like other activists and intellectuals, Ai was prevented from leaving China in late 2010. Ai suggested that the authorities wanted to prevent him from attending the ceremony in December 2010 to award the 2010 Nobel Peace Prize to fellow dissident Liu Xiaobo.[41] Ai said that he had not been invited to the ceremony, and was attempting to travel to South Korea for a meeting when he was told that he could not leave for reasons of national security.[42]

In the evening of 11 January 2011, Ai’s studio was demolished in a surprise move by the local government.[43][44]

On 3 April 2011, Ai was arrested at Beijing Capital International Airport just before catching a flight to Hong Kong and his studio facilities were searched.[45] A police contingent of approximately 50 officers came to his studio, threw a cordon around it and searched the premises. They took away laptops and the hard drive from the main computer; along with Ai, police also detained eight staff members and Ai’s wife, Lu Qing. Police also visited the mother of Ai’s two-year-old son.[46] While state media originally reported on 6 April that Ai was arrested at the airport because “his departure procedures were incomplete,”[47] the Chinese Ministry of Foreign Affairs said on 7 April that Ai was arrested under investigation for alleged economic crimes.[48] Then, on 8 April, police returned to Ai’s workshop to examine his financial affairs.[49] On 9 April, Ai’s accountant, as well as studio partner Liu Zhenggang and driver Zhang Jingsong, disappeared,[50] while Ai’s assistant Wen Tao has remained missing since Ai’s arrest on 3 April.[51] Ai’s wife said that she was summoned by the Beijing Chaoyang district tax bureau, where she was interrogated about his studio’s tax on 12 April.[52] South China Morning Post reports that Ai received at least two visits from the police, the last being on 31 March three days before his detention apparently with offers of membership to the Chinese People’s Political Consultative Conference. A staff member recalled that Ai had mentioned receiving the offer earlier, “[but Ai] didn’t say if it was a membership of the CPPCC at the municipal or national level, how he responded or whether he accepted it or not.”[52]

On 24 February, amid an online campaign for Middle East-style protests in major Chinese cities by overseas dissidents, Ai posted on his Twitter account: “I didnt care about jasmine at first, but people who are scared by jasmine sent out information about how harmful jasmine is often, which makes me realize that jasmine is what scares them the most. What a jasmine!”[53][54]

Analysts and other activists said Ai had been widely thought to be untouchable, but Nicholas Bequelin from Human Rights Watch suggested that his arrest, calculated to send the message that no one would be immune, must have had the approval of someone in the top leadership.[55] International governments, human rights groups and art institutions, among others, called for Ai’s release, while Chinese officials did not notify Ai’s family of his whereabouts.[56]

State media started describing Ai as a “deviant and a plagiarist” in early 2011.[57] The China Daily subsidiary, the Global Times editorial on 6 April 2011 attacked Ai, saying “Ai Weiwei likes to do something ‘others dare not do.’ He has been close to the red line of Chinese law. Objectively speaking, Chinese society does not have much experience in dealing with such persons. However, as long as Ai Weiwei continuously marches forward, he will inevitably touch the red line one day.”[58] Two days later, the journal scorned Western media for questioning Ai’s charge as a “catch-all crime”, and denounced the use of his political activism as a “legal shield” against everyday crimes. It said “Ai’s detention is one of the many judicial cases handled in China every day. It is pure fantasy to conclude that Ai’s case will be handled specially and unfairly.”[59] Frank Ching expressed in the South China Morning Post that how the Global Times could radically shift its position from one-day to the next was reminiscent of Alice in Wonderland.[60]

Michael Sheridan of The Times suggested that Ai had offered himself to the authorities on a platter with some of his provocative art, particularly photographs of himself nude with only a toy alpaca hiding his modesty with a caption (“grass mud horse covering the middle”). The term possesses a double meaning in Chinese: one possible interpretation was given by Sheridan as: “Fuck your mother, the party central committee”.[61]

Ming Pao in Hong Kong reacted strongly to the state media’s character attack on Ai, saying that authorities had employed “a chain of actions outside the law, doing further damage to an already weak system of laws, and to the overall image of the country.”[57] Pro-Beijing newspaper in Hong Kong, Wen Wei Po, announced that Ai was under arrest for tax evasion, bigamy and spreading indecent images on the internet, and vilified him with multiple instances of strong rhetoric.[62][63] Supporters said “the article should be seen as a mainland media commentary attacking Ai, rather than as an accurate account of the investigation.”[64]

The United States and European Union protested Ai’s detention.[65] The international arts community also mobilised petitions calling for the release of Ai: “1001 Chairs for Ai Weiwei” was organized by Creative Time of New York that calls for artists to bring chairs to Chinese embassies and consulates around the world on 17 April 2011, at 1pm local time “to sit peacefully in support of the artist’s immediate release.”[66][67] Artists in Hong Kong,[68] Germany[68] and Taiwan demonstrated and called for Ai to be released.[69]

One of the major protests by U.S. museums took place on 19 and 20 May when the Museum of Contemporary Art San Diego organized a 24-hour silent protest in which volunteer participants, including community members, media, and museum staff, occupied two traditionally styled Chinese chairs for one-hour periods.[70] The 24-hour sit-in referenced Ai’s sculpture series, Marble Chair, two of which were on view and were subsequently acquired for the Museum’s permanent collection.

The Solomon R. Guggenheim Foundation and the International Council of Museums, which organised petitions, said they had collected more than 90,000 signatures calling for the release of Ai.[71] On 13 April 2011, a group of European intellectuals led by Vclav Havel had issued an open letter to Wen Jiabao, condemning the arrest and demanding the immediate release of Ai. The signatories include Ivan Klma, Ji Grua, Jchym Topol, Elfriede Jelinek, Adam Michnik, Adam Zagajewski, Helmuth Frauendorfer; Bei Ling (Chinese:), a Chinese poet in exile drafted and also signed the open letter.[72]

On 16 May 2011, the Chinese authorities allowed Ai’s wife to visit him briefly. Liu Xiaoyuan, his attorney and personal friend, reported that Wei was in good physical condition and receiving treatment for his chronic diabetes and hypertension; he was not in a prison or hospital but under some form of house arrest.[73]

He is the subject of the 2012 documentary film Ai Weiwei: Never Sorry, directed by American filmmaker Alison Klayman, which received a special jury prize at the 2012 Sundance Film Festival and opened the Hot Docs Canadian International Documentary Festival, North America’s largest documentary festival, in Toronto on 26 April 2012.[74]

On 22 June 2011, the Chinese authorities released Ai from jail after almost three months’ detention on charges of tax evasion.[75] Beijing Fa Ke Cultural Development Ltd. (Chinese: ), a company Ai controlled, had allegedly evaded taxes and intentionally destroyed accounting documents. State media also reports that Ai was granted bail on account of Ai’s “good attitude in confessing his crimes”, willingness to pay back taxes, and his chronic illnesses.[76] According to the Chinese Foreign Ministry, he is prohibited from leaving Beijing without permission for one year.[77][78] Ai’s supporters widely viewed his detention as retaliation for his vocal criticism of the government.[79] On 23 June 2011, professor Wang Yujin of China University of Political Science and Law stated that the release of Ai on bail shows that the Chinese government could not find any solid evidence of Ai’s alleged “economic crime”.[80] On 24 June 2011, Ai told a Radio Free Asia reporter that he was thankful for the support of the Hong Kong public, and praised Hong Kong’s conscious society. Ai also mentioned that his detention by the Chinese regime was hellish (Chinese: ), and stressed that he is forbidden to say too much to reporters.[81]

After his release, his sister gave some details about his detention condition to the press, explaining that he was subjected to a kind of psychological torture: he was detained in a tiny room with constant light, and two guards were set very close to him at all times, and watched him constantly.[82] In November, Chinese authorities were again investigating Ai and his associates, this time under the charge of spreading pornography.[83][84] Lu was subsequently questioned by police, and released after several hours though the exact charges remain unclear.[85][86] In January 2012, in its International Review issue Art in America magazine featured an interview with Ai Weiwei at his home in China. J.J. Camille (the pen name of a Chinese-born writer living in New York), “neither a journalist nor an activist but simply an art lover who wanted to talk to him” had travelled to Beijing the previous September to conduct the interview and to write about his visit to “China’s most famous dissident artist” for the magazine.[87]

On 21 June 2012, Ai’s bail was lifted. Although he is allowed to leave Beijing, the police informed him that he is still prohibited from traveling to other countries because he is “suspected of other crimes,” including pornography, bigamy and illicit exchange of foreign currency.[88][89] Until 2015, he remained under heavy surveillance and restrictions of movement, but continues to criticize through his work.[90][91] In July 2015, he was given a passport and may travel abroad.[92]

In June 2011, the Beijing Local Taxation Bureau demanded a total of over 12 million yuan (US$1.85million) from Beijing Fa Ke Cultural Development Ltd. in unpaid taxes and fines,[93][94] and accorded three days to appeal the demand in writing. According to Ai’s wife, Beijing Fa Ke Cultural Development Ltd. has hired two Beijing lawyers as defense attorneys. Ai’s family state that Ai is “neither the chief executive nor the legal representative of the design company, which is registered in his wife’s name.”

Offers of donations poured in from Ai’s fans across the world when the fine was announced. Eventually an online loan campaign was initiated on 4 November 2011, and close to 9 million RMB was collected within ten days, from 30,000 contributions. Notes were folded into paper planes and thrown over the studio walls, and donations were made in symbolic amounts such as 8964 (4 June 1989, Tiananmen Massacre) or 512 (12 May 2008, Sichuan earthquake). To thank creditors and acknowledge the contributions as loans, Ai designed and issued loan receipts to all who participated in the campaign.[95] Funds raised from the campaign were used as collateral, required by law for an appeal on the tax case. Lawyers acting for Ai submitted an appeal against the fine in January 2012; the Chinese government subsequently agreed to conduct a review.[96]

In June 2012, the court heard the tax appeal case. Ai’s wife, Lu Qing, the legal representative of the design company, attended the hearing. Lu was accompanied by several lawyers and an accountant, but the witnesses they had requested to testify, including Ai, were prevented from attending a court hearing.[97] Ai asserts that the entire matter including the 81 days he spent in jail in 2011 is intended to suppress his provocations. Ai said he had no illusions as to how the case would turn out, as he believes the court will protect the government’s own interests. On 20 June, hundreds of Ai’s supporters gathered outside the Chaoyang District Court in Beijing despite a small army of police officers, some of whom videotaped the crowd and led several people away.[98] On 20 July, Ai’s tax appeal was rejected in court.[99][100] The same day Ai’s studio released “The Fake Case” which tracks the status and history of this case including a timeline and the release of official documents.[101] On 27 September, the court upheld the 2.4million tax evasion fine.[102] Ai had previously deposited 1.33million in a government-controlled account in order to appeal. Ai said he will not pay the remainder because he does not recognize the charge.[103]

In October 2012, authorities revoked the license of Beijing Fa Ke Cultural Development Ltd. for failing to re-register, an annual requirement by the administration. The company was not able to complete this procedure as its materials and stamps were confiscated by the government.[104]

On 26 April 2014, Ai’s name was removed from a group show taking place at the Shanghai Power Station of Art. The exhibition was held to celebrate the fifteenth anniversary of the art prize created by Uli Sigg in 1998, with the purpose of promoting and developing Chinese contemporary art. Ai won the Lifetime Contribution Award in 2008 and was part of the jury during the first three editions of the prize.[105] He was then invited to take part in the group show together with the other selected Chinese artists. Shortly before the exhibition’s opening, some museum workers removed his name from the list of winners and jury members painted on a wall. Also, Ai’s works Sunflower Seeds and Stools were removed from the show and kept in a museum office (see photo on Ai Weiwei’s Instagram).[106] Sigg declared that it was not his decision and that it was a decision of the Power Station of Art and the Shanghai Municipal Bureau of Culture.[105]

In May 2014, the Ullens Center for Contemporary Art, a non-profit art center situated in the 798 art district of Beijing, held a retrospective exhibition in honor of the late curator and scholar, Hans Van Dijk. Ai, a good friend of Hans and a fellow co-founder of the China Art Archives and Warehouse (CAAW), participated in the exhibition with three artworks.[107] On the day of the opening, Ai realized his name was omitted from both Chinese and English versions of the exhibition’s press release. Ai’s assistants went to the art center and removed his works.[108] It is Ai’s belief that, in omitting his name, the museum altered the historical record of van Dijk’s work with him. Ai started his own research about what actually happened, and between 23 and 25 May he interviewed the UCCA’s director, Philip Tinari, the guest curator of the exhibition, Marianne Brouwer, and the UCCA chief, Xue Mei.[107] He published the transcripts of the interviews on Instagram.[109][110][111][112][113][114][115][116][117] In one of the interviews, the CEO of the UCCA, Xue Mei, admitted that, due to the sensitive time of the exhibition, Ai’s name was taken out of the press releases on the day of the opening and it was supposed to be restored afterwards. This was to avoid problems with the Chinese authorities, who threatened to arrest her.[107]

Beijing video works

From 2003 to 2005, Ai Weiwei recorded the results of Beijings developing urban infrastructure and its social conditions.

2003, Video, 150 hours

Beginning under the Dabeiyao highway interchange, the vehicle from which Beijing 2003 was shot traveled every road within the Fourth Ring Road of Beijing and documented the road conditions. Approximately 2400 kilometers and 150 hours of footage later, it ended where it began under the Dabeiyao highway interchange. The documentation of these winding alleyways of the city center now largely torn down for redevelopment preserved a visual record of the city that is free of aesthetic judgment.

2004, Video, 10h 13m

Moving from east to west, Changan Boulevard traverses Beijings most iconic avenue. Along the boulevards 45-kilometer length, it recorded the changing densities of its far-flung suburbs, central business districts, and political core. At each 50-meter increment, the artist records a single frame for one minute. The work reveals the rhythm of Beijing as a capital city, its social structure, cityscape, socialist-planned economy, capitalist market, political power center, commercial buildings, and industrial units as pieces of a multi-layered urban collage.

2005, Video, 1h 6m

2005 Video, 1h 50m

Beijing: The Second Ring and Beijing: The Third Ring capture two opposite views of traffic flow on every bridge of each Ring Road, the innermost arterial highways of Beijing. The artist records a single frame for one minute for each view on the bridge. Beijing: The Second Ring was entirely shot on cloudy days, while the segments for Beijing: The Third Ring were entirely shot on sunny days. The films document the historic aspects and modern development of a city with a population of nearly 11 million people.

2007, video, 2h 32m[118]

This video is about Ai Weiwei’s project Fairytale for Europes most innovative five-year art event Documenta 12 in Kassel, Germany in 2007: Ai Weiwei invited 1001 Chinese citizens of different ages and from various backgrounds to Germany to experience their own fairytale for 28 days.[119] The 152 minutes film documents the whole process beginning with project preparations, over the challenge that the participants had to face until the actual travel to Germany, as well as the artists ideas behind the work. This is a work I emotionally relate to. It grows and it surprised me Ai Weiwei in Fairytale.

2008, video, 1h 18m[120]

On 15 December 2008, a citizens investigation began with the goal of seeking an explanation for the casualties of the Sichuan earthquake that happened on 12 May 2008. The investigation covered 14 counties and 74 townships within the disaster zone, and studied the conditions of 153 schools that were affected by the earthquake. By gathering and confirming comprehensive details about the students, such as their age, region, school, and grade, the group managed to affirm that there were 5,192 students who perished in the disaster. Among a hundred volunteers, 38 of them participated in fieldwork, with 25 of them being controlled by the Sichuan police for a total of 45 times. This documentary is a structural element of the citizens investigation.

2009, looped video, 1h 27m[121]

At 14:28 on 12 May 2008, an 8.0-magnitude earthquake happened in Sichuan, China. Over 5,000 students in primary and secondary schools perished in the earthquake, yet their names went unannounced. In reaction to the governments lack of transparency, a citizens investigation was initiated to find out their names and details about their schools and families. As of 2 September 2009, there were 4,851 confirmed. This video is a tribute to these perished students and a memorial for innocent lives lost.

2009, video, 48m[122]

This video documents the story of Chinese citizen Feng Zhenghu and his struggles to return home. The Shanghai authorities rejected Feng Zhenghu, originated from Wenzhou, Zhejiang, China, from returning to the country for a total of eight times in 2009. On 4 November 2009, Feng Zhenghu attempted to return home for the ninth time but the police from Shanghai used violence and kidnapped him to board a flight to Japan. Feng refused to enter Japan and decided to live in the Immigration Hall at Terminal 1 of the Narita Airport in Tokyo, as an act of protest. He relied on food gifts from tourists for sustenance and lived at a passageway in the Narita Airport for 92 days. He posted updates over Twitter, they attracted much concern and led to wide media coverage from Chinese netizens and international communities. On 31 January, Feng announced an end to his protest at the Narita Airport. On 12 February, Feng was allowed entry to China, where he reunited with his family at home in Shanghai. Ai Weiwei and his assistant Gao Yuan, went from Beijing to interview Feng Zhenghu three times at the Narita Airport of Japan on 16 November 20 November 2009 and 31 January 2010, and documented his life at the airport passageway and the entire process of his return to China. No country should refuse entry to its own citizens.

2009, video, 1h 19m[123]

Ai Weiwei studio production Laoma Tihua is a documentary of an incident during Tan Zuorens trial on 12 August 2009. Tan Zuoren was charged with inciting subversion of state power. Chengdu police detained witnessed during the trial of the civil rights advocate, which is an obstruction of justice and violence. Tan Zuoren was charged as a result of his research and questioning regarding the 5.12 Wenchuan students casualties and the corruption resulting poor building construction. Tan Zuoren was sentenced five years to prison.

2010, video, 3h[124]

In June 2008, Yang Jia carried a knife, a hammer, a gas mask, pepper spray, gloves and Molotov cocktails to the Zhabei Public Security Branch Bureau and killed six police officers, injuring another police officer and a guard. He was arrested on the scene, and was subsequently charged with intentional homicide. In the following six months, while Yang Jia was detained and trials were held, his mother has mysteriously disappeared. This video is a documentary that traces the reasons and motivations behind the tragedy and investigates into a trial process filled with shady cover-ups and questionable decisions. The film provides a glimpse into the realities of a government-controlled judicial system and its impact on the citizens lives.

2010, video, 2h 6m[125]

The future dictionary definition of crackdown will be: First cover ones head up firmly, and then beat him or her up violently. @aiww In the summer of 2010, the Chinese government began a crackdown on dissent, and Hua Hao Yue Yuan documents the stories of Liu Dejun and Liu Shasha, whose activism and outspoken attitude led them to violent abuse from the authorities. On separate occasions, they were kidnapped, beaten and thrown into remote locations. The incidents attracted much concern over the Internet, as well as wide speculation and theories about what exactly happened. This documentary presents interviews of the two victims, witnesses and concerned netizens. In which it gathers various perspectives about the two beatings, and brings us closer to the brutal reality of Chinas crackdown on crime.

2010, voice recording, 3h 41m[126]

On 24 April 2010 at 00:51, Ai Weiwei (@aiww) started a Twitter campaign to commemorate students who perished in the earthquake in Sichuan on 12 May 2008. 3,444 friends from the Internet delivered voice recordings, the names of 5,205 perished were recited 12,140 times. Remembrance is an audio work dedicated to the young people who lost their lives in the Sichuan earthquake. It expresses thoughts for the passing of innocent lives and indignation for the cover-ups on truths about sub-standard architecture, which led to the large number of schools that collapsed during the earthquake.

2010, video, 1h 8m[127]

The shooting and editing of this video lasted nearly seven months at the Ai Weiwei studio. It began near the end of 2007 in an interception organized by cat-saving volunteers in Tianjin, and the film locations included Tianjin, Shanghai, Rugao of Jiangsu, Chaoshan of Guangzhou, and Hebei Province. The documentary depicts a complete picture of a chain in the cat-trading industry. Since the end of 2009 when the government began soliciting expert opinion for the Animal Protection Act, the focus of public debate has always been on whether one should be eating cats or not, or whether cat-eating is a Chinese tradition or not. There are even people who would go as far as to say that the call to stop eating cat meat is “imposing the will of the minority on the majority”. Yet the “majority” does not understand the complete truth of cat-meat trading chains: cat theft, cat trafficking, killing cats, selling cats, and eating cats, all the various stages of the trade and how they are distributed across the country, in cities such as Beijing, Tianjin, Shanghai, Nanjing, Suzhou, Wuxi, Rugao, Wuhan, Guangzhou, and Hebei. This well-organized, smooth-running industry chain of cat abuse, cat killing and skinning has already existed among ordinary Chinese folks for 20 years, or perhaps even longer. The degree of civilization of a country can be seen from its attitude towards animals.

2011, video, 1h 1m[128]

This documentary is about the construction project curated by Herzog & de Meuron and Ai Weiwei. One hundred architects from 27 countries were chosen to participate and design a 1000 square meter villa to be built in a new community in Inner Mongolia. The 100 villas would be designed to fit a master plan designed by Ai Weiwei. On 25 January 2008, the 100 architects gathered in Ordos for a first site visit. The film Ordos 100 documents the total of three site visits to Ordos, during which time the master plan and design of each villa was completed. As of 2016, the Ordos 100 project remains unrealized.

2011, video, 54m[129]

As a sequel to Ai Weiweis film Lao Ma Ti Hua, the film So Sorry (named after the artists 2009 exhibition in Munich, Germany) shows the beginnings of the tension between Ai Weiwei and the Chinese Government. In Lao Ma Ti Hua, Ai Weiwei travels to Chengdu, Sichuan to attend the trial of the civil rights advocate Tan Zuoren, as a witness. In So Sorry, you see the investigation led by Ai Weiwei studio to identify the students who died during the Sichuan earthquake as a result of corruption and poor building constructions leading to the confrontation between Ai Weiwei and the Chengdu police. After being beaten by the police, Ai Weiwei traveled to Munich, Germany to prepare his exhibition at the museum Haus der Kunst. The result of his beating led to intense headaches caused by a brain hemorrhage and was treated by emergency surgery. These events mark the beginning of Ai Weiweis struggle and surveillance at the hands of the state police.

2011, video, 2h 22m[130]

This documentary investigates the death of popular Zhaiqiao village leader Qian Yunhui in the fishing village of Yueqing, Zhejiang province. When the local government confiscated marshlands in order to convert them into construction land, the villagers were deprived of the opportunity to cultivate these lands and be fully self-subsistent. Qian Yunhui, unafraid of speaking up for his villagers, travelled to Beijing several times to report this injustice to the central government. In order to silence him, he was detained by local government repeatedly. On 25 December 2010, Qian Yunhui was hit by a truck and died on the scene. News of the incident and photos of the scene quickly spread over the internet. The local government claimed that Qian Yunhui was the victim of an ordinary traffic accident. This film is an investigation conducted by Ai Weiwei studio into the circumstances of the incident and its connection to the land dispute case, mainly based on interviews of family members, villagers and officials. It is an attempt by Ai Weiwei to establish the facts and find out what really happened on 25 December 2010. During shooting and production, Ai Weiwei studio experienced significant obstruction and resistance from local government. The film crew was followed, sometimes physically stopped from shooting certain scenes and there were even attempts to buy off footage. All villagers interviewed for the purposes of this documentary have been interrogated or illegally detained by local government to some extent.

2011, video, 1h 1m[131]

Early in 2008, the district government of Jiading, Shanghai invited Ai Weiwei to build a studio in Malu Township, as a part of the local government’s efforts in developing its cultural assets. By August 2010, the Ai Weiwei Shanghai Studio completed all of its construction work. In October 2010, the Shanghai government declared the Ai Weiwei Shanghai Studio an illegal construction, and was subjected to demolition. On 7 November 2010, when Ai Weiwei was placed under house arrest by public security in Beijing, over 1,000 netizens attended the “River Crab Feast” at the Shanghai Studio. On 11 January 2011, the Shanghai city government forcibly demolished the Ai Weiwei Studio within a day, without any prior notice.

2013, video, 1h 17m[132]

This video tells the story of Liu Ximei, who at her birth in 1985 was given to relatives to be raised because she was born in violation of Chinas strict one-child policy. When she was ten years old, Liu was severely injured while working in the fields and lost large amounts of blood. While undergoing treatment at a local hospital, she was given a blood transfusion that was later revealed to be contaminated with HIV. Following this exposure to the virus, Liu contracted AIDS. According to official statistics, in 2001 there were 850,000 AIDS sufferers in China, many of whom contracted the illness in the 1980s and 1990s as the result of a widespread plasma market operating in rural, impoverished areas and using unsafe collection methods.

2014, video, 2h 8m[133]

Ai Weiweis Appeal 15,220,910.50 opens with Ai Weiweis mother at the Venice Biennial in the summer of 2013 examining Ais large S.A.C.R.E.D. installation portraying his 81-day imprisonment. The documentary goes onto chronologically reconstruct the events that occurred from the time he was arrested at the Beijing airport in April 2011 to his final court appeal in September 2012. The film portrays the day-to-day activity surrounding Ai Weiwei, his family and his associates ranging from consistent visits by the authorities, interviews with reporters, support and donations from fans, and court dates. The Film premiered at the International Film Festival Rotterdam on 23 January 2014.

2015, video, 30m[134]

This documentary on the Fukushima Art Project is about artist Ai Weiweis investigation of the site as well as the project’s installation process. In August 2014, Ai Weiwei was invited as one of the participating artists for the Fukushima Nuclear Zone by the Japanese art coalition ChimPom, as part of the project Dont Follow the Wind . Ai accepted the invitation and sent his assistant Ma Yan to the exclusion zone in Japan to investigate the site. The Fukushima Nuclear Exclusion Zone is thus far located within the 20-kilometer radius of land area of the Fukushima Daiichi Nuclear Power Plant. 25,000 people have already been evacuated from the Exclusion Zone. Both water and electric circuits were cut off. Entrance restriction is expected to be relieved in the next thirty years, or even longer. The art project will also be open to public at that time. The three spots usable as exhibition spaces by the artists are all former residential houses, among which exhibition site one and two were used for working and lodging; and exhibition site three was used as a community entertainment facility with an ostrich farm. Ai brought about two projects, “A Ray of Hope” and “Family Album” after analyzing materials and information generated from the site. In “A Ray of Hope”, a solar photovoltaic system is built on exhibition site one, on the second level of the old warehouse. Integral LED lighting devices are used in the two rooms. The lights would turn on automatically from 7 to 10pm, and from 6 to 8am daily. This lighting system is the only light source in the Exclusion Zone after this project was installed. Photos of Ai and his studio staff at Caochangdi that make up project “Family Album” are displayed on exhibition site two and three, in the seven rooms where locals used to live. The twenty-two selected photos are divided in five categories according to types of event spanning eight years. Among these photos, six of them were taken from the site investigation at the 2008 Sichuan earthquake; two were taken during the time when he was illegally detained after pleading the Tan Zuoren case in Chengdu, China in August 2009; and three others taken during his surgical treatment for his head injury from being attacked in the head by police officers in Chengdu; five taken of him being followed by the police and his Beijing studio Fake Design under surveillance due to the studio tax case from 2011 to 2012; four are photos of Ai Weiwei and his family from year 2011 to year 2013; and the other two were taken earlier of him in his studio in Caochangdi (One taken in 2005 and the other in 2006).

A feature documentary directed and co-produced by Ai Weiwei about the global refugee crisis.

Ai’s visual art includes sculptural installations, woodworking, video and photography. “Ai Weiwei: According to What,” adapted and expanded by the Hirshhorn Museum and Sculpture Garden from a 2009 exhibition at Tokyo’s Mori Art Museum, was Ai’s first North American museum retrospective. [135] It opened at the Hirshhorn in Washington, D.C. in 2013, and subsequently traveled to the Brooklyn Museum, New York, [136] and two other venues. His works address his investigation into the aftermath of the Sichuan earthquake and responses to the Chinese government’s detention and surveillance of him. [137] His recent public pieces have called attention to the Syrian refugee crisis.[138]

(1995) Performance in which Ai lets an ancient ceramic urn fall from his hands and smash to pieces on the ground. The performance was memorialized in a series of three photographic still frames.[139]

(2008) Sculpture resembling a park bench or tree trunk, but its cross-section is a map of China. It is four metres long and weighs 635 kilograms. It is made from wood salvaged from Qing Dynasty temples.[140]

(2008) Ming dynasty table cut in half and rejoined at a right angle to rest two feet on the wall and two on the floor. The reconstruction was completed using Chinese period specific joinery techniques.[141]

(2008-2012) 150 tons of twisted steel reinforcements recovered from the 2008 Sichuan earthquake building collapse sites were straightened out and displayed as an installation.[142]

(2010) Opening in October 2010 at the Tate Museum in London, Ai displayed 100 million handmade and painted porcelain sunflower seeds. These seeds weight about 150 tons and were made over a span of two and a half years by 1,600 Jingdezhen artisans. This city made porcelain for the government for over one thousand years. The artisans produced the sunflower seeds in the traditional method that the city is known for, in which a thirty step procedure is employed. The sculpture relates back to Chairman Mao’s rule and the Chinese Communist Party. The combination of all the seeds represent that together, the people of China can stand up and overthrow the Chinese Communist Party. Along with this, the seeds represent China’s growing mass production stemming from the consumerist culture in the west. The sculpture directly challenges the Made in China mantra that China is known for, considering the labor-intensive and traditional method of creating the work.[143]

(2010) Sculptures in marble to resemble the cameras placed in front of Ai’s studio.[144]

(2011) Sculptures of zodiac animals inspired by the water clock-fountain at the Old Summer Palace.[145]

(2014) Han dynasty vase with the Coca-Cola logo brushed on in red acrylic paint.[146]

(2014) 32 Qing dynasty stools joined together in a cluster with legs pointing out.[147]

(2014) Individual porcelain ornaments, each painted with characters for “free speech”, which when set together form a map of China.[148]

(2014) Consisting of 176 2D-portraits in Lego which are set onto a large floor space, Trace was commissioned by the FOR-SITE Foundation, the United States National Park Service and the Golden Gate Park Conservancy. The original installation was at Alcatraz Prison in San Francisco Bay; the 176 portraits being of various political prisoners and prisoners of conscience. After seeing one million visitors during its one-year display at Alcatraz, the installation was moved and put on display at the Hirshhorn Museum in Washington, D.C. (in a modified form; the pieces had to be arranged to fit the circular floor space). The display at the Hirshhorn ran from June 28, 2017 January 1, 2018. The display also included two versions of his wallpaper work The Animal That Looks Like a Llama but Is Really an Alpaca and a video running on a loop.[149]

(2017) As the culmination of Ai’s experiences visiting 40 refugee camps in 2016, Law of the Journey featured an all-black, 230-foot-long inflatable boat carrying 258 faceless refugee figures. The art piece is currently on display at the National Gallery in Prague until January 7, 2018.[150]

(2017) Permanent exhibit, unique setting of two Iron Trees from now on frame the Shrine of the Book in Jerusalem, Israel where Dead Sea Scrolls are preserved[151][152]

(2017) On the view in Israel Museum until the end of October 2017, Journey of Laziz is a video installation, showing mental breakdown and overall suffering of tiger living in the “world’s worst ZOO” in Gaza[151][152]

(2017) On view at the Park Avenue Armory through August 6, 2017, Hansel and Gretel is an installation exploring the theme of surveillance. The project, a collaboration of Ai Weiwei and architects Jacques Herzog and Pierre de Meuron, features surveillance cameras equipped with facial recognition software, near-infrared floor projections, tethered, autonomous drones and sonar beacons. A companion website includes a curatorial statement, artist biographies, a livestream of the installation and a timeline of surveillance technology from ancient to modern times.[153]

More:

Ai Weiwei – Wikipedia

Posted in Ai

Wit.ai

Wit.ai makes it easy for developers to build applications and devices that you can talk or text to. Our vision is to empower developers with an open and extensible natural language platform. Wit.ai learns human language from every interaction, and leverages the community: what’s learned is shared across developers.

The rest is here:

Wit.ai

Posted in Ai

AI File – What is it and how do I open it?

Did your computer fail to open an AI file? We explain what AI files are and recommend software that we know can open or convert your AI files.

AI is the acronym for Adobe Illustrator. Files that have the .ai extension are drawing files that the Adobe Illustrator application has created.

The Adobe Illustrator application was developed by Adobe Systems. The files created by this application are composed of paths that are connected by points and are saved in vector format. The technology used to create these files allows the user to re-size the AI image without losing any of the image’s quality.

Some third-party programs allow users to “rastersize” the images created in Adobe Illustrator, which allows them to convert the AI file into bitmap format. While this may make the file size smaller and easier to open across multiple applications, some of the file quality may be lost in the process.

Read the rest here:

AI File – What is it and how do I open it?

Posted in Ai

Artificial intelligence – Wikipedia

Artificial intelligence (AI, also machine intelligence, MI) is intelligence demonstrated by machines, in contrast to the natural intelligence (NI) displayed by humans and other animals. In computer science AI research is defined as the study of “intelligent agents”: any device that perceives its environment and takes actions that maximize its chance of successfully achieving its goals.[1] Colloquially, the term “artificial intelligence” is applied when a machine mimics “cognitive” functions that humans associate with other human minds, such as “learning” and “problem solving”. See glossary of artificial intelligence.

The scope of AI is disputed: as machines become increasingly capable, tasks considered as requiring “intelligence” are often removed from the definition, a phenomenon known as the AI effect, leading to the quip “AI is whatever hasn’t been done yet.”[3] For instance, optical character recognition is frequently excluded from “artificial intelligence”, having become a routine technology.[4] Capabilities generally classified as AI as of 2017[update] include successfully understanding human speech, competing at the highest level in strategic game systems (such as chess and Go[6]), autonomous cars, intelligent routing in content delivery networks, military simulations, and interpreting complex data, including images and videos.

Artificial intelligence was founded as an academic discipline in 1956, and in the years since has experienced several waves of optimism,[7][8] followed by disappointment and the loss of funding (known as an “AI winter”),[9][10] followed by new approaches, success and renewed funding.[8][11] For most of its history, AI research has been divided into subfields that often fail to communicate with each other.[12] These sub-fields are based on technical considerations, such as particular goals (e.g. “robotics” or “machine learning”),[13] the use of particular tools (“logic” or “neural networks”), or deep philosophical differences.[14][15][16] Subfields have also been based on social factors (particular institutions or the work of particular researchers).[12]

The traditional problems (or goals) of AI research include reasoning, knowledge, planning, learning, natural language processing, perception and the ability to move and manipulate objects.[13] General intelligence is among the field’s long-term goals.[17] Approaches include statistical methods, computational intelligence, and traditional symbolic AI. Many tools are used in AI, including versions of search and mathematical optimization, neural networks and methods based on statistics, probability and economics. The AI field draws upon computer science, mathematics, psychology, linguistics, philosophy and many others.

The field was founded on the claim that human intelligence “can be so precisely described that a machine can be made to simulate it”.[18] This raises philosophical arguments about the nature of the mind and the ethics of creating artificial beings endowed with human-like intelligence, issues which have been explored by myth, fiction and philosophy since antiquity.[19] Some people also consider AI to be a danger to humanity if it progresses unabatedly.[20] Others believe that AI, unlike previous technological revolutions, will create a risk of mass unemployment.[21]

In the twenty-first century, AI techniques have experienced a resurgence following concurrent advances in computer power, large amounts of data, and theoretical understanding; and AI techniques have become an essential part of the technology industry, helping to solve many challenging problems in computer science.[22][11]

While thought-capable artificial beings appeared as storytelling devices in antiquity,[23] the idea of actually trying to build a machine to perform useful reasoning may have begun with Ramon Llull (c. 1300 CE). With his Calculus ratiocinator, Gottfried Leibniz extended the concept of the calculating machine (Wilhelm Schickard engineered the first one around 1623), intending to perform operations on concepts rather than numbers. Since the 19th century, artificial beings are common in fiction, as in Mary Shelley’s Frankenstein or Karel apek’s R.U.R. (Rossum’s Universal Robots).[25]

The study of mechanical or “formal” reasoning began with philosophers and mathematicians in antiquity. The study of mathematical logic led directly to Alan Turing’s theory of computation, which suggested that a machine, by shuffling symbols as simple as “0” and “1”, could simulate any conceivable act of mathematical deduction. This insight, that digital computers can simulate any process of formal reasoning, is known as the ChurchTuring thesis.[26][pageneeded] Along with concurrent discoveries in neurobiology, information theory and cybernetics, this led researchers to consider the possibility of building an electronic brain.[27] The first work that is now generally recognized as AI was McCullouch and Pitts’ 1943 formal design for Turing-complete “artificial neurons”.

The field of AI research was born at a workshop at Dartmouth College in 1956.[28] Attendees Allen Newell (CMU), Herbert Simon (CMU), John McCarthy (MIT), Marvin Minsky (MIT) and Arthur Samuel (IBM) became the founders and leaders of AI research.[29] They and their students produced programs that the press described as “astonishing”: computers were learning checkers strategies (c. 1954)[31] (and by 1959 were reportedly playing better than the average human),[32] solving word problems in algebra, proving logical theorems (Logic Theorist, first run c. 1956) and speaking English.[33] By the middle of the 1960s, research in the U.S. was heavily funded by the Department of Defense[34] and laboratories had been established around the world.[35] AI’s founders were optimistic about the future: Herbert Simon predicted, “machines will be capable, within twenty years, of doing any work a man can do”. Marvin Minsky agreed, writing, “within a generation… the problem of creating ‘artificial intelligence’ will substantially be solved”.[7]

They failed to recognize the difficulty of some of the remaining tasks. Progress slowed and in 1974, in response to the criticism of Sir James Lighthill and ongoing pressure from the US Congress to fund more productive projects, both the U.S. and British governments cut off exploratory research in AI. The next few years would later be called an “AI winter”,[9] a period when obtaining funding for AI projects was difficult.

In the early 1980s, AI research was revived by the commercial success of expert systems,[37] a form of AI program that simulated the knowledge and analytical skills of human experts. By 1985 the market for AI had reached over a billion dollars. At the same time, Japan’s fifth generation computer project inspired the U.S and British governments to restore funding for academic research.[8] However, beginning with the collapse of the Lisp Machine market in 1987, AI once again fell into disrepute, and a second, longer-lasting hiatus began.[10]

In the late 1990s and early 21st century, AI began to be used for logistics, data mining, medical diagnosis and other areas.[22] The success was due to increasing computational power (see Moore’s law), greater emphasis on solving specific problems, new ties between AI and other fields and a commitment by researchers to mathematical methods and scientific standards.[38] Deep Blue became the first computer chess-playing system to beat a reigning world chess champion, Garry Kasparov on 11 May 1997.

Advanced statistical techniques (loosely known as deep learning), access to large amounts of data and faster computers enabled advances in machine learning and perception.[citation needed] By the mid 2010s, machine learning applications were used throughout the world.[citation needed] In a Jeopardy! quiz show exhibition match, IBM’s question answering system, Watson, defeated the two greatest Jeopardy champions, Brad Rutter and Ken Jennings, by a significant margin. The Kinect, which provides a 3D bodymotion interface for the Xbox 360 and the Xbox One use algorithms that emerged from lengthy AI research[41] as do intelligent personal assistants in smartphones.[42] In March 2016, AlphaGo won 4 out of 5 games of Go in a match with Go champion Lee Sedol, becoming the first computer Go-playing system to beat a professional Go player without handicaps.[6][43] In the 2017 Future of Go Summit, AlphaGo won a three-game match with Ke Jie,[44] who at the time continuously held the world No. 1 ranking for two years.[45][46] This marked the completion of a significant milestone in the development of Artificial Intelligence as Go is an extremely complex game, more so than Chess.

According to Bloomberg’s Jack Clark, 2015 was a landmark year for artificial intelligence, with the number of software projects that use AI within Google increased from a “sporadic usage” in 2012 to more than 2,700 projects. Clark also presents factual data indicating that error rates in image processing tasks have fallen significantly since 2011.[47] He attributes this to an increase in affordable neural networks, due to a rise in cloud computing infrastructure and to an increase in research tools and datasets.[11] Other cited examples include Microsoft’s development of a Skype system that can automatically translate from one language to another and Facebook’s system that can describe images to blind people.[47]

A typical AI perceives its environment and takes actions that maximize its chance of successfully achieving its goals.[1] An AI’s intended goal function can be simple (“1 if the AI wins a game of Go, 0 otherwise”) or complex (“Do actions mathematically similar to the actions that got you rewards in the past”). Goals can be explicitly defined, or can be induced. If the AI is programmed for “reinforcement learning”, goals can be implicitly induced by rewarding some types of behavior and punishing others.[a] Alternatively, an evolutionary system can induce goals by using a “fitness function” to mutate and preferentially replicate high-scoring AI systems; this is similar to how animals evolved to innately desire certain goals such as finding food, or how dogs can be bred via artificial selection to possess desired traits. Some AI systems, such as nearest-neighbor, instead reason by analogy; these systems are not generally given goals, except to the degree that goals are somehow implicit in their training data. Such systems can still be benchmarked if the non-goal system is framed as a system whose “goal” is to successfully accomplish its narrow classification task.[50]

AI often revolves around the use of algorithms. An algorithm is a set of unambiguous instructions that a mechanical computer can execute.[b] A complex algorithm is often built on top of other, simpler, algorithms. A simple example of an algorithm is the following recipe for optimal play at tic-tac-toe:

Many AI algorithms are capable of learning from data; they can enhance themselves by learning new heuristics (strategies, or “rules of thumb”, that have worked well in the past), or can themselves write other algorithms. Some of the “learners” described below, including Bayesian networks, decision trees, and nearest-neighbor, could theoretically, if given infinite data, time, and memory, learn to approximate any function, including whatever combination of mathematical functions would best describe the entire world. These learners could therefore, in theory, derive all possible knowledge, by considering every possible hypothesis and matching it against the data. In practice, it is almost never possible to consider every possibility, because of the phenomenon of “combinatorial explosion”, where the amount of time needed to solve a problem grows exponentially. Much of AI research involves figuring out how to identify and avoid considering broad swaths of possibililities that are unlikely to be fruitful.[52] For example, when viewing a map and looking for the shortest driving route from Denver to New York in the East, one can in most cases skip looking at any path through San Francisco or other areas far to the West; thus, an AI wielding an pathfinding algorithm like A* can avoid the combinatorial explosion that would ensue if every possible route had to be ponderously considered in turn.[54]

Learners work on the basis that strategies, algorithms, and inferences that worked well in the past are likely to continue working well in the future. These inferences can be obvious, such as “since the sun rose every morning for the last 10,000 days, it will probably rise tomorrow morning as well”. They can be nuanced, such as “X% of families have geographically separate species with color variants, so there is an Y% chance that undiscovered black swans exist”. Learners also work on the basis of “Occam’s razor”: The simplest theory that explains the data is the likeliest. Therefore, to be successful, a learner must be designed such that it prefers simpler theories to complex theories, except in cases where the complex theory is proven substantially better. Settling on a bad, overly complex theory gerrymandered to fit all the past training data is known as overfitting. Many systems attempt to reduce overfitting by rewarding a theory in accordance with how well it fits the data, but penalizing the theory in accordance with how complex the theory is. Besides classic overfitting, learners can also disappoint by “learning the wrong lesson”. A toy example is that an image classifier trained only on pictures of brown horses and black cats might conclude that all brown patches are likely to be horses. A real-world example is that, unlike humans, current image classifiers don’t determine the spatial relationship between components of the picture; instead, they learn abstract patterns of pixels that humans are oblivious to, but that linearly correlate with images of certain types of real objects. Faintly superimposing such a pattern on a legitimate image results in an “adversarial” image that the system misclassifies.[c][57][58][59]

The overall research goal of artificial intelligence is to create technology that allows computers and machines to function in an intelligent manner. The general problem of simulating (or creating) intelligence has been broken down into sub-problems. These consist of particular traits or capabilities that researchers expect an intelligent system to display. The traits described below have received the most attention.[13]

Early researchers developed algorithms that imitated step-by-step reasoning that humans use when they solve puzzles or make logical deductions.[60] By the late 1980s and 1990s, AI research had developed methods for dealing with uncertain or incomplete information, employing concepts from probability and economics.[61]

Human beings ordinarily use fast, intuitive judgments rather than step-by-step deduction that early AI research was able to model.[62] AI has progressed using “sub-symbolic” problem solving: embodied agent approaches emphasize the importance of sensorimotor skills to higher reasoning; neural net research attempts to simulate the structures inside the brain that give rise to this skill; statistical approaches to AI mimic the human ability to guess.

Knowledge representation[63] and knowledge engineering[64] are central to AI research. Many of the problems machines are expected to solve will require extensive knowledge about the world. Among the things that AI needs to represent are: objects, properties, categories and relations between objects;[65] situations, events, states and time;[66] causes and effects;[67] knowledge about knowledge (what we know about what other people know);[68] and many other, less well researched domains. A representation of “what exists” is an ontology: the set of objects, relations, concepts, and properties formally described so that software agents can interpret them. The semantics of these are captured as description logic concepts, roles, and individuals, and typically implemented as classes, properties, and individuals in the Web Ontology Language.[69] The most general ontologies are called upper ontologies, which attempt to provide a foundation for all other knowledge[70] by acting as mediators between domain ontologies that cover specific knowledge about a particular knowledge domain (field of interest or area of concern). Such formal knowledge representations are suitable for content-based indexing and retrieval, scene interpretation, clinical decision support, knowledge discovery via automated reasoning (inferring new statements based on explicitly stated knowledge), etc. Video events are often represented as SWRL rules, which can be used, among others, to automatically generate subtitles for constrained videos.[71]

Among the most difficult problems in knowledge representation are:

Intelligent agents must be able to set goals and achieve them.[78] They need a way to visualize the futurea representation of the state of the world and be able to make predictions about how their actions will change itand be able to make choices that maximize the utility (or “value”) of available choices.[79]

In classical planning problems, the agent can assume that it is the only system acting in the world, allowing the agent to be certain of the consequences of its actions.[80] However, if the agent is not the only actor, then it requires that the agent can reason under uncertainty. This calls for an agent that can not only assess its environment and make predictions, but also evaluate its predictions and adapt based on its assessment.[81]

Multi-agent planning uses the cooperation and competition of many agents to achieve a given goal. Emergent behavior such as this is used by evolutionary algorithms and swarm intelligence.[82]

Machine learning, a fundamental concept of AI research since the field’s inception,[83] is the study of computer algorithms that improve automatically through experience.[84][85]

Unsupervised learning is the ability to find patterns in a stream of input. Supervised learning includes both classification and numerical regression. Classification is used to determine what category something belongs in, after seeing a number of examples of things from several categories. Regression is the attempt to produce a function that describes the relationship between inputs and outputs and predicts how the outputs should change as the inputs change. In reinforcement learning[86] the agent is rewarded for good responses and punished for bad ones. The agent uses this sequence of rewards and punishments to form a strategy for operating in its problem space. These three types of learning can be analyzed in terms of decision theory, using concepts like utility. The mathematical analysis of machine learning algorithms and their performance is a branch of theoretical computer science known as computational learning theory.[citation needed]

Within developmental robotics, developmental learning approaches are elaborated upon to allow robots to accumulate repertoires of novel skills through autonomous self-exploration, social interaction with human teachers, and the use of guidance mechanisms (active learning, maturation, motor synergies, etc.).[87][88]

Natural language processing[91] gives machines the ability to read and understand human language. A sufficiently powerful natural language processing system would enable natural language user interfaces and the acquisition of knowledge directly from human-written sources, such as newswire texts. Some straightforward applications of natural language processing include information retrieval, text mining, question answering[92] and machine translation.[93]

A common method of processing and extracting meaning from natural language is through semantic indexing. Although these indexes require a large volume of user input, it is expected that increases in processor speeds and decreases in data storage costs will result in greater efficiency.

Machine perception[94] is the ability to use input from sensors (such as cameras, microphones, tactile sensors, sonar and others) to deduce aspects of the world. Computer vision[95] is the ability to analyze visual input. A few selected subproblems are speech recognition,[96] facial recognition and object recognition.[97]

The field of robotics[98] is closely related to AI. Intelligence is required for robots to handle tasks such as object manipulation[99] and navigation, with sub-problems such as localization, mapping, and motion planning. These systems require that an agent is able to: Be spatially cognizant of its surroundings, learn from and build a map of its environment, figure out how to get from one point in space to another, and execute that movement (which often involves compliant motion, a process where movement requires maintaining physical contact with an object).[101]

Affective computing is the study and development of systems that can recognize, interpret, process, and simulate human affects. It is an interdisciplinary field spanning computer sciences, psychology, and cognitive science. While the origins of the field may be traced as far back as the early philosophical inquiries into emotion, the more modern branch of computer science originated with Rosalind Picard’s 1995 paper on “affective computing”.[108][109] A motivation for the research is the ability to simulate empathy, where the machine would be able to interpret human emotions and adapts its behavior to give an appropriate response to those emotions.

Emotion and social skills[110] are important to an intelligent agent for two reasons. First, being able to predict the actions of others by understanding their motives and emotional states allow an agent to make better decisions. Concepts such as game theory, decision theory, necessitate that an agent be able to detect and model human emotions. Second, in an effort to facilitate humancomputer interaction, an intelligent machine may want to display emotions (even if it does not experience those emotions itself) to appear more sensitive to the emotional dynamics of human interaction.

A sub-field of AI addresses creativity both theoretically (the philosophical psychological perspective) and practically (the specific implementation of systems that generate novel and useful outputs).

Many researchers think that their work will eventually be incorporated into a machine with artificial general intelligence, combining all the skills mentioned above and even exceeding human ability in most or all these areas.[17][111] A few believe that anthropomorphic features like artificial consciousness or an artificial brain may be required for such a project.[112][113]

Many of the problems above also require that general intelligence be solved. For example, even specific straightforward tasks, like machine translation, require that a machine read and write in both languages (NLP), follow the author’s argument (reason), know what is being talked about (knowledge), and faithfully reproduce the author’s original intent (social intelligence). A problem like machine translation is considered “AI-complete”, but all of these problems need to be solved simultaneously in order to reach human-level machine performance.

There is no established unifying theory or paradigm that guides AI research. Researchers disagree about many issues.[114] A few of the most long standing questions that have remained unanswered are these: should artificial intelligence simulate natural intelligence by studying psychology or neurobiology? Or is human biology as irrelevant to AI research as bird biology is to aeronautical engineering?[14] Can intelligent behavior be described using simple, elegant principles (such as logic or optimization)? Or does it necessarily require solving a large number of completely unrelated problems?[15] Can intelligence be reproduced using high-level symbols, similar to words and ideas? Or does it require “sub-symbolic” processing?[16] John Haugeland, who coined the term GOFAI (Good Old-Fashioned Artificial Intelligence), also proposed that AI should more properly be referred to as synthetic intelligence,[115] a term which has since been adopted by some non-GOFAI researchers.

Stuart Shapiro divides AI research into three approaches, which he calls computational psychology, computational philosophy, and computer science. Computational psychology is used to make computer programs that mimic human behavior.[118] Computational philosophy, is used to develop an adaptive, free-flowing computer mind.[118] Implementing computer science serves the goal of creating computers that can perform tasks that only people could previously accomplish.[118] Together, the humanesque behavior, mind, and actions make up artificial intelligence.

In the 1940s and 1950s, a number of researchers explored the connection between neurobiology, information theory, and cybernetics. Some of them built machines that used electronic networks to exhibit rudimentary intelligence, such as W. Grey Walter’s turtles and the Johns Hopkins Beast. Many of these researchers gathered for meetings of the Teleological Society at Princeton University and the Ratio Club in England.[27] By 1960, this approach was largely abandoned, although elements of it would be revived in the 1980s.

When access to digital computers became possible in the middle 1950s, AI research began to explore the possibility that human intelligence could be reduced to symbol manipulation. The research was centered in three institutions: Carnegie Mellon University, Stanford and MIT, and each one developed its own style of research. John Haugeland named these approaches to AI “good old fashioned AI” or “GOFAI”.[119] During the 1960s, symbolic approaches had achieved great success at simulating high-level thinking in small demonstration programs. Approaches based on cybernetics or neural networks were abandoned or pushed into the background.[120] Researchers in the 1960s and the 1970s were convinced that symbolic approaches would eventually succeed in creating a machine with artificial general intelligence and considered this the goal of their field.

Economist Herbert Simon and Allen Newell studied human problem-solving skills and attempted to formalize them, and their work laid the foundations of the field of artificial intelligence, as well as cognitive science, operations research and management science. Their research team used the results of psychological experiments to develop programs that simulated the techniques that people used to solve problems. This tradition, centered at Carnegie Mellon University would eventually culminate in the development of the Soar architecture in the middle 1980s.[121][122]

Unlike Newell and Simon, John McCarthy felt that machines did not need to simulate human thought, but should instead try to find the essence of abstract reasoning and problem solving, regardless of whether people used the same algorithms.[14] His laboratory at Stanford (SAIL) focused on using formal logic to solve a wide variety of problems, including knowledge representation, planning and learning.[123] Logic was also the focus of the work at the University of Edinburgh and elsewhere in Europe which led to the development of the programming language Prolog and the science of logic programming.[124]

Researchers at MIT (such as Marvin Minsky and Seymour Papert)[125] found that solving difficult problems in vision and natural language processing required ad-hoc solutions they argued that there was no simple and general principle (like logic) that would capture all the aspects of intelligent behavior. Roger Schank described their “anti-logic” approaches as “scruffy” (as opposed to the “neat” paradigms at CMU and Stanford).[15] Commonsense knowledge bases (such as Doug Lenat’s Cyc) are an example of “scruffy” AI, since they must be built by hand, one complicated concept at a time.[126]

When computers with large memories became available around 1970, researchers from all three traditions began to build knowledge into AI applications.[127] This “knowledge revolution” led to the development and deployment of expert systems (introduced by Edward Feigenbaum), the first truly successful form of AI software.[37] The knowledge revolution was also driven by the realization that enormous amounts of knowledge would be required by many simple AI applications.

By the 1980s progress in symbolic AI seemed to stall and many believed that symbolic systems would never be able to imitate all the processes of human cognition, especially perception, robotics, learning and pattern recognition. A number of researchers began to look into “sub-symbolic” approaches to specific AI problems.[16] Sub-symbolic methods manage to approach intelligence without specific representations of knowledge.

This includes embodied, situated, behavior-based, and nouvelle AI. Researchers from the related field of robotics, such as Rodney Brooks, rejected symbolic AI and focused on the basic engineering problems that would allow robots to move and survive.[128] Their work revived the non-symbolic viewpoint of the early cybernetics researchers of the 1950s and reintroduced the use of control theory in AI. This coincided with the development of the embodied mind thesis in the related field of cognitive science: the idea that aspects of the body (such as movement, perception and visualization) are required for higher intelligence.

Interest in neural networks and “connectionism” was revived by David Rumelhart and others in the middle of the 1980s.[129] Neural networks are an example of soft computing — they are solutions to problems which cannot be solved with complete logical certainty, and where an approximate solution is often sufficient. Other soft computing approaches to AI include fuzzy systems, evolutionary computation and many statistical tools. The application of soft computing to AI is studied collectively by the emerging discipline of computational intelligence.[130]

In the 1990s, AI researchers developed sophisticated mathematical tools to solve specific subproblems. These tools are truly scientific, in the sense that their results are both measurable and verifiable, and they have been responsible for many of AI’s recent successes. The shared mathematical language has also permitted a high level of collaboration with more established fields (like mathematics, economics or operations research). Stuart Russell and Peter Norvig describe this movement as nothing less than a “revolution” and “the victory of the neats”.[38] Critics argue that these techniques (with few exceptions) are too focused on particular problems and have failed to address the long-term goal of general intelligence. There is an ongoing debate about the relevance and validity of statistical approaches in AI, exemplified in part by exchanges between Peter Norvig and Noam Chomsky.

In the course of 60 or so years of research, AI has developed a large number of tools to solve the most difficult problems in computer science. A few of the most general of these methods are discussed below.

Many problems in AI can be solved in theory by intelligently searching through many possible solutions:[138] Reasoning can be reduced to performing a search. For example, logical proof can be viewed as searching for a path that leads from premises to conclusions, where each step is the application of an inference rule.[139] Planning algorithms search through trees of goals and subgoals, attempting to find a path to a target goal, a process called means-ends analysis.[140] Robotics algorithms for moving limbs and grasping objects use local searches in configuration space.[99] Many learning algorithms use search algorithms based on optimization.

Simple exhaustive searches[141] are rarely sufficient for most real world problems: the search space (the number of places to search) quickly grows to astronomical numbers. The result is a search that is too slow or never completes. The solution, for many problems, is to use “heuristics” or “rules of thumb” that prioritize choices in favor of those that are more likely to reach a goal, and to do so in a shorter number of steps. In some search methodologies heuristics can also serve to entirely eliminate some choices that are unlikely to lead to a goal (called “pruning the search tree”). Heuristics supply the program with a “best guess” for the path on which the solution lies.[142] Heuristics limit the search for solutions into a smaller sample size.

A very different kind of search came to prominence in the 1990s, based on the mathematical theory of optimization. For many problems, it is possible to begin the search with some form of a guess and then refine the guess incrementally until no more refinements can be made. These algorithms can be visualized as blind hill climbing: we begin the search at a random point on the landscape, and then, by jumps or steps, we keep moving our guess uphill, until we reach the top. Other optimization algorithms are simulated annealing, beam search and random optimization.[143]

Evolutionary computation uses a form of optimization search. For example, they may begin with a population of organisms (the guesses) and then allow them to mutate and recombine, selecting only the fittest to survive each generation (refining the guesses). Forms of evolutionary computation include swarm intelligence algorithms (such as ant colony or particle swarm optimization)[144] and evolutionary algorithms (such as genetic algorithms, gene expression programming, and genetic programming).[145]

Logic[146] is used for knowledge representation and problem solving, but it can be applied to other problems as well. For example, the satplan algorithm uses logic for planning[147] and inductive logic programming is a method for learning.[148]

Several different forms of logic are used in AI research. Propositional or sentential logic[149] is the logic of statements which can be true or false. First-order logic[150] also allows the use of quantifiers and predicates, and can express facts about objects, their properties, and their relations with each other. Fuzzy logic,[151] is a version of first-order logic which allows the truth of a statement to be represented as a value between 0 and 1, rather than simply True (1) or False (0). Fuzzy systems can be used for uncertain reasoning and have been widely used in modern industrial and consumer product control systems. Subjective logic[citation needed] models uncertainty in a different and more explicit manner than fuzzy-logic: a given binomial opinion satisfies belief + disbelief + uncertainty = 1 within a Beta distribution. By this method, ignorance can be distinguished from probabilistic statements that an agent makes with high confidence.

Default logics, non-monotonic logics and circumscription[73] are forms of logic designed to help with default reasoning and the qualification problem. Several extensions of logic have been designed to handle specific domains of knowledge, such as: description logics;[65] situation calculus, event calculus and fluent calculus (for representing events and time);[66] causal calculus;[67] belief calculus;[152] and modal logics.[68]

Many problems in AI (in reasoning, planning, learning, perception and robotics) require the agent to operate with incomplete or uncertain information. AI researchers have devised a number of powerful tools to solve these problems using methods from probability theory and economics.[153]

Bayesian networks[154] are a very general tool that can be used for a large number of problems: reasoning (using the Bayesian inference algorithm),[155] learning (using the expectation-maximization algorithm),[d][157] planning (using decision networks)[158] and perception (using dynamic Bayesian networks).[159] Bayesian networks are used in AdSense to choose what ads to place and on XBox Live to rate and match players. Probabilistic algorithms can also be used for filtering, prediction, smoothing and finding explanations for streams of data, helping perception systems to analyze processes that occur over time (e.g., hidden Markov models or Kalman filters).[159]

A key concept from the science of economics is “utility”: a measure of how valuable something is to an intelligent agent. Precise mathematical tools have been developed that analyze how an agent can make choices and plan, using decision theory, decision analysis,[161] and information value theory.[79] These tools include models such as Markov decision processes,[162] dynamic decision networks,[159] game theory and mechanism design.[163]

The simplest AI applications can be divided into two types: classifiers (“if shiny then diamond”) and controllers (“if shiny then pick up”). Controllers do, however, also classify conditions before inferring actions, and therefore classification forms a central part of many AI systems. Classifiers are functions that use pattern matching to determine a closest match. They can be tuned according to examples, making them very attractive for use in AI. These examples are known as observations or patterns. In supervised learning, each pattern belongs to a certain predefined class. A class can be seen as a decision that has to be made. All the observations combined with their class labels are known as a data set. When a new observation is received, that observation is classified based on previous experience.[164]

A classifier can be trained in various ways; there are many statistical and machine learning approaches. The decision tree[165] is perhaps the most widely used machine learning algorithm. Other widely used classifiers are the neural network,[167] k-nearest neighbor algorithm,[e][169] kernel methods such as the support vector machine (SVM),[f][171] Gaussian mixture model[172] and the extremely popular naive Bayes classifier.[g][174] The performance of these classifiers have been compared over a wide range of tasks. Classifier performance depends greatly on the characteristics of the data to be classified. There is no single classifier that works best on all given problems; this is also referred to as the “no free lunch” theorem. Determining a suitable classifier for a given problem is still more an art than science.[175]

Neural networks, or neural nets, were inspired by the architecture of neurons in the human brain. A simple “neuron” N accepts input from multiple other neurons, each of which, when activated (or “fired”), cast a weighted “vote” for or against whether neuron N should itself activate. Learning requires an algorithm to adjust these weights based on the training data; one simple algorithm (dubbed “fire together, wire together”) is to increase the weight between two connected neurons when the activation of one triggers the successful activation of another. The net forms “concepts” that are distributed among a subnetwork of shared[h] neurons that tend to fire together; a concept meaning “leg” might be coupled with a subnetwork meaning “foot” that includes the sound for “foot”. Neurons have a continuous spectrum of activation; in addition, neurons can process inputs in a nonlinear way rather than weighing straightforward votes. Modern neural nets can learn both continuous functions and, surprisingly, digital logical operations. Neural networks’ early successes included predicting the stock market and (in 1995) a mostly self-driving car.[i] In the 2010s, advances in neural networks using deep learning thrust AI into widespread public consciousness and contributed to an enormous upshift in corporate AI spending; for example, AI-related M&A in 2017 was over 25 times as large as in 2015.[177][178]

The study of non-learning artificial neural networks[167] began in the decade before the field of AI research was founded, in the work of Walter Pitts and Warren McCullouch. Frank Rosenblatt invented the perceptron, a learning network with a single layer, similar to the old concept of linear regression. Early pioneers also include Alexey Grigorevich Ivakhnenko, Teuvo Kohonen, Stephen Grossberg, Kunihiko Fukushima, Christoph von der Malsburg, David Willshaw, Shun-Ichi Amari, Bernard Widrow, John Hopfield, Eduardo R. Caianiello, and others.

The main categories of networks are acyclic or feedforward neural networks (where the signal passes in only one direction) and recurrent neural networks (which allow feedback and short-term memories of previous input events). Among the most popular feedforward networks are perceptrons, multi-layer perceptrons and radial basis networks.[179] Neural networks can be applied to the problem of intelligent control (for robotics) or learning, using such techniques as Hebbian learning (“fire together, wire together”), GMDH or competitive learning.[180]

Today, neural networks are often trained by the backpropagation algorithm, which had been around since 1970 as the reverse mode of automatic differentiation published by Seppo Linnainmaa,[181][182] and was introduced to neural networks by Paul Werbos.[183][184][185]

Hierarchical temporal memory is an approach that models some of the structural and algorithmic properties of the neocortex.[186]

In short, most neural networks use some form of gradient descent on a hand-created neural topology. However, some research groups, such as Uber, argue that simple neuroevolution to mutate new neural network topologies and weights may be competitive with sophisticated gradient descent approaches. One advantage of neuroevolution is that it may be less prone to get caught in “dead ends”.[187]

Deep learning is any artificial neural network that can learn a long chain of causal links. For example, a feedforward network with six hidden layers can learn a seven-link causal chain (six hidden layers + output layer) and has a “credit assignment path” (CAP) depth of seven. Many deep learning systems need to be able to learn chains ten or more causal links in length.[188] Deep learning has transformed many important subfields of artificial intelligence, including computer vision, speech recognition, natural language processing and others.[189][190][188]

According to one overview,[191] the expression “Deep Learning” was introduced to the Machine Learning community by Rina Dechter in 1986[192] and gained traction after Igor Aizenberg and colleagues introduced it to Artificial Neural Networks in 2000.[193] The first functional Deep Learning networks were published by Alexey Grigorevich Ivakhnenko and V. G. Lapa in 1965.[194][pageneeded] These networks are trained one layer at a time. Ivakhnenko’s 1971 paper[195] describes the learning of a deep feedforward multilayer perceptron with eight layers, already much deeper than many later networks. In 2006, a publication by Geoffrey Hinton and Ruslan Salakhutdinov introduced another way of pre-training many-layered feedforward neural networks (FNNs) one layer at a time, treating each layer in turn as an unsupervised restricted Boltzmann machine, then using supervised backpropagation for fine-tuning. Similar to shallow artificial neural networks, deep neural networks can model complex non-linear relationships. Over the last few years, advances in both machine learning algorithms and computer hardware have led to more efficient methods for training deep neural networks that contain many layers of non-linear hidden units and a very large output layer.[197]

Deep learning often uses convolutional neural networks (CNNs), whose origins can be traced back to the Neocognitron introduced by Kunihiko Fukushima in 1980.[198] In 1989, Yann LeCun and colleagues applied backpropagation to such an architecture. In the early 2000s, in an industrial application CNNs already processed an estimated 10% to 20% of all the checks written in the US.[199] Since 2011, fast implementations of CNNs on GPUs have won many visual pattern recognition competitions.[188]

CNNs with 12 convolutional layers were used in conjunction with reinforcement learning by Deepmind’s “AlphaGo Lee”, the program that beat a top Go champion in 2016.[200]

Early on, deep learning was also applied to sequence learning with recurrent neural networks (RNNs)[201] which are in theory Turing complete[202] and can run arbitrary programs to process arbitrary sequences of inputs. The depth of an RNN is unlimited and depends on the length of its input sequence; thus, an RNN is an example of deep learning.[188] RNNs can be trained by gradient descent[203][204][205] but suffer from the vanishing gradient problem.[189][206] In 1992, it was shown that unsupervised pre-training of a stack of recurrent neural networks can speed up subsequent supervised learning of deep sequential problems.[207]

Numerous researchers now use variants of a deep learning recurrent NN called the long short-term memory (LSTM) network published by Hochreiter & Schmidhuber in 1997.[208] LSTM is often trained by Connectionist Temporal Classification (CTC).[209] At Google, Microsoft and Baidu this approach has revolutionised speech recognition.[210][211][212] For example, in 2015, Google’s speech recognition experienced a dramatic performance jump of 49% through CTC-trained LSTM, which is now available through Google Voice to billions of smartphone users.[213] Google also used LSTM to improve machine translation,[214] Language Modeling[215] and Multilingual Language Processing.[216] LSTM combined with CNNs also improved automatic image captioning[217] and a plethora of other applications.

Early symbolic AI inspired Lisp[218] and Prolog,[219] which dominated early AI programming. Modern AI development often uses mainstream languages such as Python or C++,[220] or niche languages such as Wolfram Language.[221]

In 1950, Alan Turing proposed a general procedure to test the intelligence of an agent now known as the Turing test. This procedure allows almost all the major problems of artificial intelligence to be tested. However, it is a very difficult challenge and at present all agents fail.[222]

Artificial intelligence can also be evaluated on specific problems such as small problems in chemistry, hand-writing recognition and game-playing. Such tests have been termed subject matter expert Turing tests. Smaller problems provide more achievable goals and there are an ever-increasing number of positive results.[citation needed]

For example, performance at draughts (i.e. checkers) is optimal,[citation needed] performance at chess is high-human and nearing super-human (see computer chess:computers versus human) and performance at many everyday tasks (such as recognizing a face or crossing a room without bumping into something) is sub-human.

A quite different approach measures machine intelligence through tests which are developed from mathematical definitions of intelligence. Examples of these kinds of tests start in the late nineties devising intelligence tests using notions from Kolmogorov complexity and data compression.[223] Two major advantages of mathematical definitions are their applicability to nonhuman intelligences and their absence of a requirement for human testers.

A derivative of the Turing test is the Completely Automated Public Turing test to tell Computers and Humans Apart (CAPTCHA). As the name implies, this helps to determine that a user is an actual person and not a computer posing as a human. In contrast to the standard Turing test, CAPTCHA is administered by a machine and targeted to a human as opposed to being administered by a human and targeted to a machine. A computer asks a user to complete a simple test then generates a grade for that test. Computers are unable to solve the problem, so correct solutions are deemed to be the result of a person taking the test. A common type of CAPTCHA is the test that requires the typing of distorted letters, numbers or symbols that appear in an image undecipherable by a computer.

AI is relevant to any intellectual task. Modern artificial intelligence techniques are pervasive and are too numerous to list here. Frequently, when a technique reaches mainstream use, it is no longer considered artificial intelligence; this phenomenon is described as the AI effect.

High-profile examples of AI include autonomous vehicles (such as drones and self-driving cars), medical diagnosis, creating art (such as poetry), proving mathematical theorems, playing games (such as Chess or Go), search engines (such as Google search), online assistants (such as Siri), image recognition in photographs, spam filtering, prediction of judicial decisions[227] and targeting online advertisements.[228][229]

With social media sites overtaking TV as a source for news for young people and news organisations increasingly reliant on social media platforms for generating distribution,[230] major publishers now use artificial intelligence (AI) technology to post stories more effectively and generate higher volumes of traffic.[231]

There are a number of competitions and prizes to promote research in artificial intelligence. The main areas promoted are: general machine intelligence, conversational behavior, data-mining, robotic cars, robot soccer and games.

Artificial intelligence is breaking into the healthcare industry by assisting doctors. According to Bloomberg Technology, Microsoft has developed AI to help doctors find the right treatments for cancer.[232] There is a great amount of research and drugs developed relating to cancer. In detail, there are more than 800 medicines and vaccines to treat cancer. This negatively affects the doctors, because there are too many options to choose from, making it more difficult to choose the right drugs for the patients. Microsoft is working on a project to develop a machine called “Hanover”. Its goal is to memorize all the papers necessary to cancer and help predict which combinations of drugs will be most effective for each patient. One project that is being worked on at the moment is fighting myeloid leukemia, a fatal cancer where the treatment has not improved in decades. Another study was reported to have found that artificial intelligence was as good as trained doctors in identifying skin cancers.[233] Another study is using artificial intelligence to try and monitor multiple high-risk patients, and this is done by asking each patient numerous questions based on data acquired from live doctor to patient interactions.[234]

According to CNN, there was a recent study by surgeons at the Children’s National Medical Center in Washington which successfully demonstrated surgery with an autonomous robot. The team supervised the robot while it performed soft-tissue surgery, stitching together a pig’s bowel during open surgery, and doing so better than a human surgeon, the team claimed.[235] IBM has created its own artificial intelligence computer, the IBM Watson, which has beaten human intelligence (at some levels). Watson not only won at the game show Jeopardy! against former champions,[236] but, was declared a hero after successfully diagnosing a women who was suffering from leukemia.[237]

Advancements in AI have contributed to the growth of the automotive industry through the creation and evolution of self-driving vehicles. As of 2016, there are over 30 companies utilizing AI into the creation of driverless cars. A few companies involved with AI include Tesla, Google, and Apple.[238]

Many components contribute to the functioning of self-driving cars. These vehicles incorporate systems such as braking, lane changing, collision prevention, navigation and mapping. Together, these systems, as well as high performance computers, are integrated into one complex vehicle.[239]

Recent developments in autonomous automobiles have made the innovation of self-driving trucks possible, though they are still in the testing phase. The UK government has passed legislation to begin testing of self-driving truck platoons in 2018.[240] Self-driving truck platoons are a fleet of self-driving trucks following the lead of one non-self-driving truck, so the truck platoons aren’t entirely autonomous yet. Meanwhile, the Daimler, a German automobile corporation, is testing the Freightliner Inspiration which is a semi-autonomous truck that will only be used on the highway.[241]

One main factor that influences the ability for a driver-less automobile to function is mapping. In general, the vehicle would be pre-programmed with a map of the area being driven. This map would include data on the approximations of street light and curb heights in order for the vehicle to be aware of its surroundings. However, Google has been working on an algorithm with the purpose of eliminating the need for pre-programmed maps and instead, creating a device that would be able to adjust to a variety of new surroundings.[242] Some self-driving cars are not equipped with steering wheels or brake pedals, so there has also been research focused on creating an algorithm that is capable of maintaining a safe environment for the passengers in the vehicle through awareness of speed and driving conditions.[243]

Another factor that is influencing the ability for a driver-less automobile is the safety of the passenger. To make a driver-less automobile, engineers must program it to handle high risk situations. These situations could include a head on collision with pedestrians. The car’s main goal should be to make a decision that would avoid hitting the pedestrians and saving the passengers in the car. But there is a possibility the car would need to make a decision that would put someone in danger. In other words, the car would need to decide to save the pedestrians or the passengers.[244] The programing of the car in these situations is crucial to a successful driver-less automobile.

Financial institutions have long used artificial neural network systems to detect charges or claims outside of the norm, flagging these for human investigation. The use of AI in banking can be traced back to 1987 when Security Pacific National Bank in US set-up a Fraud Prevention Task force to counter the unauthorised use of debit cards. Programs like Kasisto and Moneystream are using AI in financial services.

Banks use artificial intelligence systems today to organize operations, maintain book-keeping, invest in stocks, and manage properties. AI can react to changes overnight or when business is not taking place.[245] In August 2001, robots beat humans in a simulated financial trading competition.[246] AI has also reduced fraud and financial crimes by monitoring behavioral patterns of users for any abnormal changes or anomalies.[247]

The use of AI machines in the market in applications such as online trading and decision making has changed major economic theories.[248] For example, AI based buying and selling platforms have changed the law of supply and demand in that it is now possible to easily estimate individualized demand and supply curves and thus individualized pricing. Furthermore, AI machines reduce information asymmetry in the market and thus making markets more efficient while reducing the volume of trades. Furthermore, AI in the markets limits the consequences of behavior in the markets again making markets more efficient. Other theories where AI has had impact include in rational choice, rational expectations, game theory, Lewis turning point, portfolio optimization and counterfactual thinking.

In video games, artificial intelligence is routinely used to generate dynamic purposeful behavior in non-player characters (NPCs). In addition, well-understood AI techniques are routinely used for pathfinding. Some researchers consider NPC AI in games to be a “solved problem” for most production tasks. Games with more atypical AI include the AI director of Left 4 Dead (2008) and the neuroevolutionary training of platoons in Supreme Commander 2 (2010).[249][250]

A platform (or “computing platform”) is defined as “some sort of hardware architecture or software framework (including application frameworks), that allows software to run”. As Rodney Brooks pointed out many years ago, it is not just the artificial intelligence software that defines the AI features of the platform, but rather the actual platform itself that affects the AI that results, i.e., there needs to be work in AI problems on real-world platforms rather than in isolation.

A wide variety of platforms has allowed different aspects of AI to develop, ranging from expert systems such as Cyc to deep-learning frameworks to robot platforms such as the Roomba with open interface.[252] Recent advances in deep artificial neural networks and distributed computing have led to a proliferation of software libraries, including Deeplearning4j, TensorFlow, Theano and Torch.

Collective AI is a platform architecture that combines individual AI into a collective entity, in order to achieve global results from individual behaviors.[253][254] With its collective structure, developers can crowdsource information and extend the functionality of existing AI domains on the platform for their own use, as well as continue to create and share new domains and capabilities for the wider community and greater good.[255] As developers continue to contribute, the overall platform grows more intelligent and is able to perform more requests, providing a scalable model for greater communal benefit.[254] Organizations like SoundHound Inc. and the Harvard John A. Paulson School of Engineering and Applied Sciences have used this collaborative AI model.[256][254]

A McKinsey Global Institute study found a shortage of 1.5 million highly trained data and AI professionals and managers[257] and a number of private bootcamps have developed programs to meet that demand, including free programs like The Data Incubator or paid programs like General Assembly.[258]

Read more from the original source:

Artificial intelligence – Wikipedia

Posted in Ai

12345...102030...