As Mr. Eck says, these systems are at least approaching the point still many, many years away when a machine can instantly build a new Beatles song or perhaps trillions of new Beatles songs, each sounding a lot like the music the Beatles themselves recorded, but also a little different. But that end game as much a way of undermining art as creating it is not what he is after. There are so many other paths to explore beyond mere mimicry. The ultimate idea is not to replace artists but to give them tools that allow them to create in entirely new ways.

In the 1990s, at that juke joint in New Mexico, Mr. Eck combined Johnny Rotten and Johnny Cash. Now, he is building software that does much the same thing. Using neural networks, he and his team are crossbreeding sounds from very different instruments say, a bassoon and a clavichord creating instruments capable of producing sounds no one has ever heard.

Much as a neural network can learn to identify a cat by analyzing hundreds of cat photos, it can learn the musical characteristics of a bassoon by analyzing hundreds of notes. It creates a mathematical representation, or vector, that identifies a bassoon. So, Mr. Eck and his team have fed notes from hundreds of instruments into a neural network, building a vector for each one. Now, simply by moving a button across a screen, they can combine these vectors to create new instruments. One may be 47 percent bassoon and 53 percent clavichord. Another might switch the percentages. And so on.

For centuries, orchestral conductors have layered sounds from various instruments atop one other. But this is different. Rather than layering sounds, Mr. Eck and his team are combining them to form something that didnt exist before, creating new ways that artists can work. Were making the next film camera, Mr. Eck said. Were making the next electric guitar.

Called NSynth, this particular project is only just getting off the ground. But across the worlds of both art and technology, many are already developing an appetite for building new art through neural networks and other A.I. techniques. This work has exploded over the last few years, said Adam Ferris, a photographer and artist in Los Angeles. This is a totally new aesthetic.

In 2015, a separate team of researchers inside Google created DeepDream, a tool that uses neural networks to generate haunting, hallucinogenic imagescapes from existing photography, and this has spawned new art inside Google and out. If the tool analyzes a photo of a dog and finds a bit of fur that looks vaguely like an eyeball, it will enhance that bit of fur and then repeat the process. The result is a dog covered in swirling eyeballs.

At the same time, a number of artists like the well-known multimedia performance artist Trevor Paglen or the lesser-known Adam Ferris are exploring neural networks in other ways. In January, Mr. Paglen gave a performance in an old maritime warehouse in San Francisco that explored the ethics of computer vision through neural networks that can track the way we look and move. While members of the avant-garde Kronos Quartet played onstage, for example, neural networks analyzed their expressions in real time, guessing at their emotions.

The tools are new, but the attitude is not. Allison Parrish, a New York University professor who builds software that generates poetry, points out that artists have been using computers to generate art since the 1950s. Much like as Jackson Pollock figured out a new way to paint by just opening the paint can and splashing it on the canvas beneath him, she said, these new computational techniques create a broader palette for artists.

A year ago, David Ha was a trader with Goldman Sachs in Tokyo. During his lunch breaks he started toying with neural networks and posting the results to a blog under a pseudonym. Among other things, he built a neural network that learned to write its own Kanji, the logographic Chinese characters that are not so much written as drawn.

Soon, Mr. Eck and other Googlers spotted the blog, and now Mr. Ha is a researcher with Google Magenta. Through a project called SketchRNN, he is building neural networks that can draw. By analyzing thousands of digital sketches made by ordinary people, these neural networks can learn to make images of things like pigs, trucks, boats or yoga poses. They dont copy what people have drawn. They learn to draw on their own, to mathematically identify what a pig drawing looks like.

Then, you ask them to, say, draw a pig with a cats head, or to visually subtract a foot from a horse or sketch a truck that looks like a dog or build a boat from a few random squiggly lines. Next to NSynth or DeepDream, these may seem less like tools that artists will use to build new works. But if you play with them, you realize that they are themselves art, living works built by Mr. Ha. A.I. isnt just creating new kinds of art; its creating new kinds of artists.

Read more:

How AI Is Creating Building Blocks to Reshape Music and Art - New York Times

You might not have noticed, but over the weekend a little coup took place. On Friday night, in front of a crowd of thousands, an AI bot beat a professional human player at Dota 2 one of the worlds most popular video games. The human champ, the affable Danil "Dendi" Ishutin, threw in the towel after being killed three times, saying he couldnt beat the unstoppable bot. It feels a little bit like human, said Dendi. But at the same time, its something else.

The bots patron was none other than tech billionaire Elon Musk, who helped found and fund the institution that designed it, OpenAI. Musk wasnt present, but made his feelings known on Twitter, saying: OpenAI first ever to defeat world's best players in competitive eSports. Vastly more complex than traditional board games like chess & Go. Even more exciting, said OpenAI, was that the AI had taught itself everything it knew. It learned purely by playing successive versions of itself, amassing lifetimes of in-game experience over the course of just two weeks.

But how big a deal is all this? Was Friday nights showdown really more impressive than Googles AI victories at the board game Go? The short answer is probably not, but it still represents a significant step forward both for the world of e-sports and the world of artificial intelligence.

First, we need to look at Musks claim that Dota is vastly more complex than traditional board games like chess & Go. This is completely true. Real-time battle and strategy games like Dota and Starcraft II pose major challenges that computers just cant handle yet. Not only do these games demand long-term strategic thinking, but unlike board games they keep vital information hidden from players. You can see everything thats happening on a chess board, but you cant in a video game. This means you have to predict and preempt what your opponent will do. It takes imagination and intuition.

In Dota, this complexity is increased as human players are asked to work together in teams of five, coordinating strategies that will change on the fly based on which characters players choose. To make things even more complex, there are more than 100 different characters in-game, each with their own unique skill set; and characters can be equipped with a number of unique items, each of which can be game-winning if deployed at the right moment. All this means its basically impossible to comprehensively program winning strategies into a Dota bot.

But, the game that OpenAIs bot played was nowhere near as complex as all this. Instead of 5v5, it took on humans at 1v1; and instead of choosing a character, both human and computer were limited to the same hero a fellow named the Shadow Fiend, who has a pretty straightforward set of attacks. My colleague Vlad Savov, a confirmed Dota addict who also wrote up his thoughts on Fridays match, said the 1v1 match represents only a fraction of the complexity of the full team contest. So: probably not as complex as Go.

The second major caveat is knowing what advantages OpenAIs agent had over its human opponents. One of the major points of discussion in the AI community was whether or not the bot had access to Dotas bot API which would let it tap directly into streams of information from the game, like the distances between players. OpenAIs Greg Brockman confirmed to The Verge that the AI did indeed use the API, and that certain techniques were hardcoded in the agent, including the items it should use in the game. It was also taught certain strategies (like one called creep block) using a trial-and-error technique known as reinforcement learning. Basically, it did get a little coaching.

Andreas Theodorou, a games AI researcher at the University of Bath and an experienced Dota player, explains why this makes a difference. One of the main things in Dota is that you need to calculate distances to know how far some [attacks] travel, he says. The API allows bots to have specific indications of range. So you can say, If someone is in 500 meters range, do that, but the human player has to calculate it themselves, learning through trial and error. It really gives them an advantage if they have access to information that a human player does not. This is particularly true in a 1v1 setting with a hero like Shadow Fiend; where players have to focus on timing their attacks correctly, rather than overall strategy.

Brockmans response is that this sort of skill is trivial for an AI to learn, and was never the focus of OpenAIs research. He says the institutes bot could have done without information from the API, but youd just be spending a lot more of your time learning to do vision, which we already know works, so whats the benefit?

So, knowing all this, should we dismiss OpenAIs victory? Not at all, says Brockman. He points out that, perhaps more important than the bots victory, was how it taught itself in the first place. While previous AI champions like AlphaGo have learned how to play games by soaking up past matches by human champions, OpenAIs bot taught itself (nearly) everything it knows.

You have this system that has just played against itself, and it has learned robust enough strategies to beat the top pros. Thats not something you should take for granted, says Brockman. And its a big question for any machine learning system: how does complexity get into the model? Where does it come from?

As OpenAIs Dota bot shows, he says, we dont have to teach computers complexity: they can learn it themselves. And although some of the bots behavior was preprogrammed, it did develop some strategies by itself. For example, it learned how to fake out its opponents by pretending to trigger an attack, only to cancel at the last second, leaving the human player to dodge an attack that never comes exactly like a feint in boxing.

Others, though, are still a little skeptical. AI researcher Denny Britz, who wrote a popular blog post that put the victory in context, tells The Verge that its difficult to judge the scale of this achievement without knowing more technical details. (Brockman says these are forthcoming, but couldnt give an exact time frame.) Its not clear what the technical contribution is at this point before the paper comes out, says Britz.

Theodorou points out that although OpenAIs bot beat Dendi onstage, once players got a good look at its tactics, they were able to outwit it. If you look at the strategies they used, they played outside the box a bit and they won, he says. The players used offbeat strategies the sort that wouldnt faze a human opponent, but which the AI had never seen before. It didnt look like the bot was flexible enough, says Theodorou. (Brockman counters that once the bot learned these strategies, it wouldnt fall for them twice.)

All the experts agree that this was a major achievement, but that the real challenge is yet to come. That will be a 5v5 match, where OpenAIs agents have to manage not just a duel in the middle of the map, but a sprawling, chaotic battlefield, with multiple heroes, dozens of support units, and unexpected twists. Brockman says that OpenAI is currently targeting next years grand Dota tournament in 12 months time to pull this off. Between now and then, theres much more training to be done.

Read this article:

Did Elon Musk's AI champ destroy humans at video games? It's complicated - The Verge

MITs Computer Science and Artificial Intelligence Lab (CSAIL) wants to ensure your streaming video experience stays smooth. A research team led by MIT professor Mohammad Alizadeh has developed an artificial intelligence (dubbed Pensieve) that can select the best algorithms for ensuring video streams both without interruption, and at the best possible playback quality.

The method improves upon existing tech, including the adaptive bitrate (ABR) method used by YouTube that throttles back quality to keep videos playing, albeit with pixelation and other artifacts. The AI can select different algorithms depending on what kind of network conditions a device is experiencing, cutting down on the downsides associated with any one method.

During experimentation, the CSAIL research team behind this method found that video streamed with between 10 and 30 percent less rebuffing, with 10 to 25 percent improved quality. Those gains would definitely add up to a significantly improved experience for most video viewers, especially over a long period.

The difference in CSAILs Pensieve approach vs. traditional methods is mainly in its use of a neural network instead of sticking to a strictly algorithmic-based approach. The neural net learns how to optimize through a reward system that incentivizes smoother video playback, rather than setting out defined rules about what algorithmic techniques to use when buffering video.

Researchers say the system is also potentially tweakable on the user end, depending on what they want to prioritize in playback: You could, for instance, set Pensieve to optimize for playback quality, or conversely, for playback speed, or even for conservation of data.

The team is making their project code open source for Pensieve at SIGCOMM next week in LA, and they expect that when trained on a larger data set, it could provide even greater improvements in terms of performance and quality. Theyre also now going to test applying it to VR video, since the high bitrates required for a quality experience there are well suited to the kinds of improvements Pensieve can offer.

See the rest here:

MIT's AI streaming software aims to stop those video stutters - TechCrunch

Xu Lis software scans more faces than maybe any on earth. He has the Chinese police to thank.

Xu runs SenseTime Group Ltd., which makes artificial intelligence software that recognizes objects and faces, and counts Chinas biggest smartphone brands as customers. In July, SenseTime raised $410 million, a sum it said was the largest single round for an AI company to date. That feat may soon be topped, probably by another startup in China.

The nation is betting heavily on AI. Money is pouring in from Chinas investors, big internet companies and its government, driven by a belief that the technology can remake entire sectors of the economy, as well as national security. A similar effort is underway in the U.S., but in this new global arms race, China has three advantages: A vast pool of engineers to write the software, a massive base of 751 million internet users to test it on, and most importantlystaunch government support that includes handing over gobs of citizens data - something that makes Western officials squirm.

Data is key because thats how AI engineers train and test algorithms to adapt and learn new skills without human programmers intervening. SenseTime built its video analysis software using footage from the police force in Guangzhou, a southern city of 14 million. Most Chinese mega-cities have set up institutes for AI that include some data-sharing arrangements, according to Xu. "In China, the population is huge, so its much easier to collect the data for whatever use-scenarios you need," he said. "When we talk about data resources, really the largest data source is the government."

This flood of data will only rise. China just enshrined the pursuit of AI into a kind of national technology constitution. A state plan, issued in July, calls for the nation to become the leader in the industry by 2030. Five years from then, the government claims the AI industry will create 400 billion yuan ($59 billion) in economic activity. Chinas tech titans, particularly Tencent Holdings Ltd. and Baidu Inc., are getting on board. And the science is showing up in unexpected places: Shanghais courts are testing an AI system that scours criminal cases to judge the validity of evidence used by all sides, ostensibly to prevent wrongful prosecutions.

Data access has always been easier in China, but now people in government, organizations and companies have recognized the value of data, said Jiebo Luo,a computer science professor at the University of Rochester who has researched China. As long as they can find someone they trust, they are willing to share it.

The AI-MATHS machine took the math portion of Chinas annual university entrance exam in Chengdu.

Photographer: AFP via Getty Images

Every major U.S. tech company is investing deeply as well. Machine learning -- a type of AI that lets driverless cars see, chatbots speak and machines parse scores of financial information -- demands computers learn from raw data instead of hand-cranked programming. Getting access to that data is a permanent slog. Chinas command-and-control economy, and its thinner privacy concerns, mean that country can dispense video footage, medical records, banking information and other wells of data almost whenever it pleases.

Xu argued this is a global phenomenon. "Theres a trend toward making data more public. For example, NHS and Google recently shared some medical image data," he said. But that example does more to illustrate Chinas edge.

DeepMind, the AI lab of Googles Alphabet Inc., has labored for nearly two years to access medical records from the U.K.s National Health Service for a diagnostics app. The agency began a trial with the company using 1.6 million patient records. Last month, the top U.K. privacy watchdog declared the trial violates British data-protection laws, throwing its future into question.

Go player Lee Se-Dol, right, in a match against Googles AlphaGo, during the DeepMind Challenge Match in March 2016.

Photographer: Google via Getty Images

Contrast that with how officials handled a project in Fuzhou. Government leaders from that southeastern Chinese city of more than seven million people held an event on June 26. Venture capital firm Sequoia Capital helped organize the event, which included representatives from Dell Inc., International Business Machines Corp. and Lenovo Group Ltd.A spokeswoman for Dell characterized the event as the nations first "Healthcare and Medical Big Data Ecology Summit."

The summit involved a vast handover of data. At the press conference, city officials shared 80 exabytes worth of heart ultrasound videos, according to one company that participated. With the massive data set, some of the companies were tasked with building an AI tool that could identify heart disease, ideally at rates above medical experts. They were asked to turn it around by the fall.

"The Chinese AI market is moving fast because people are willing to take risks and adopt new technology more quickly in a fast-growing economy," said Chris Nicholson, co-founder of Skymind Inc., one of the companies involved in the event. "AI needs big data, and Chinese regulators are now on the side of making data accessible to accelerate AI."

Representatives from IBM and Lenovo declined to comment. Last month, Lenovo Chief Executive Officer Yang Yuanqing said he will invest $1 billion into AI research over the next three to four years.

Along with health, finance can be a lucrative business in China. In part, thats because the country has far less stringent privacy regulations and concerns than the West. For decades the government has kept a secret file on nearly everyone in China called a dangan. The records run the gamut from health reports and school marks to personality assessments and club records. This dossier can often decide a citizens future -- whether they can score a promotion or be allowed to reside in the city they work.

U.S. companies that partner in China stress that AI efforts, like those in Fuzhou, are for non-military purposes. Luo, the computer science professor, said most national security research efforts are relegated to select university partners. However, one stated goal of the governments national plan is for a greater integration of civilian, academic and military development of AI.

The government also revealed in 2015 that it was building a nationwide database that would score citizens on their trustworthiness, which in turn would feed into their credit ratings. Last year, China Premier Li Keqiang said 80 percent of the nations data was in public hands and would be opened to the public, with an unspecific pledge to protect privacy. The raging popularity of live video feeds -- where Chinese internet users spend hours watching daily footage caught by surveillance video -- shows the gulf in privacy concerns between the country and the West. Embraced in China, the security cameras also reel in mountains of valuable data.

Some machine-learning researchers dispel the idea that data can be a panacea. Advanced AI operations, like DeepMind, often rely on "simulated" data, co-founder Demis Hassabis explained during a trip to China in May. DeepMind has used Atari video games to train its systems. Engineers building self-driving car software frequently test it this way, simulating stretches of highway or crashes virtually.

"Sure, there might be data sets you could get access to in China that you couldnt in the U.S.," said Oren Etzioni, director of the Allen Institute for Artificial Intelligence. "But that does not put them in a terrific position vis-a-vis AI. Its still a question of the algorithm, the insights and the research."

Historically, the country has been a lightweight in those regards. Its suffered through a "brain drain," a flight of academics and specialists out of the country. "China currently has a talent shortage when it comes to top tier AI experts," said Connie Chan, a partner at venture capital firm Andreessen Horowitz. "While there have been more deep learning papers published in China than the U.S. since 2016, those papers have not been as influential as those from the U.S. and U.K."

Exclusive insights on technology around the world.

Get Fully Charged, from Bloomberg Technology.

But China is gaining ground. The country is producing more top engineers, who craft AI algorithms for U.S. companies and, increasingly, Chinese ones. Chinese universities and private firms are actively wooing AI researchers from across the globe. Juo, the University of Rochester professor, said top researchers can get offers of $500,000 or more in annual compensation from U.S. tech companies, while Chinese companies will often double that.

Meanwhile, Chinas homegrown talent is starting to shine. A popular benchmark in AI research is the ImageNet competition, an annual challenge to devise a visual recognition system with the lowest error rate. Like last year, this years top winners were dominated by researchers from China, including a team from the Ministry of Public Securitys Third Research Institute.

Relentless pollution in metropolises like Beijing and Shanghai has hurt Chinese companies ability to nab top tech talent. In response, some are opening shop in Silicon Valley. Tencent recently set up an AI research lab in Seattle.

Photographer: David Paul Morris/Bloomberg

Baidu managed to pull a marquee name from that city. The firm recruited Qi Lu, one of Microsofts top executives, to return to China to lead the search giants push into AI. He touted the technologys potential for enhancing Chinas "national strength" and cited a figure that nearly half of the bountiful academic research on the subject globally has ethnically Chinese authors, using the Mandarin term "huaren" -- a term for ethnic Chinese that echoes government rhetoric.

"China has structural advantages, because China can acquire more and better data to power AI development," Lu told the cheering crowd of Chinese developers. "We must have the chance to lead the world!"

Continue reading here:

China's Plan for World Domination in AI Isn't So Crazy After All - Bloomberg

Out of this world

Stanford University and Intel

By Matt Reynolds

Take a look at the above image of a German street. At a glance it could be a blurry dashcam photo, or a snap thats gone through one of those apps that turns photos into paintings.

But you wont find this street anywhere on Google Maps. Thats because it was generated by an imaginative neural network, stitching together its memories of real streets it was trained on.

Nothing in the image actually exists, says Qifeng Chen at Stanford University, California, and Intel. Instead, his AI works from rough layouts that tell it what should be in each part of the image. The centre of the image might be labelled road while other sections are labelled trees or cars its painting by numbers for an AI artist.

Chen says the technique could eventually create game worlds that truly resemble the real world. Using deep learning to render video games could be the future, he says. He has already experimented with using the algorithm to replace the game world in Grand Theft Auto V.

Noah Snavely at Cornell University, New York, is impressed. Generating realistic-looking artificial scenes is a tricky problem, he says, and even the best existing approaches cant do it. Chens system creates the largest and most detailed examples of their kind he has seen.

Snavely says that the technology could allow people to describe a world, and then have an AI build it in virtual reality. Itd be great if you could conjure up a photorealistic scene just by describing it aloud, he says.

Chens system starts by processing a photo of a real street it hasnt seen before, but that has been labelled so the AI knows which bits are supposed to be cars, people, roads and so on. The AI then uses this layout as a guide to generate a completely new image.

The AI was trained on 3000 images of German streets, so when it comes across part of the photolabelled car it draws on its existing knowledge to generate a car there in its own creation. We want the network to memorise what its seen in the data, Chen says.

Intel researchers will present the work at this years International Conference on Computer Vision, which takes place in Venice, Italy, in late October.

The algorithm was also trained and tested on a smaller database of photos of domestic interiors, but Snavely says that to realise its potential it needs a data set that captures the true diversity of the world. Thats easier said than done, however, as each component in the training images needs to be labelled by hand, and creating a data set with that level of detail is extremely labour-intensive.

Chen says his system still has a long way to go before it can build truly photorealistic worlds. The images it produces right now have a blurry, dreamlike quality, as the network isnt able to fill in all the details we expect in photos. He is already working on a larger version of the system that he hopes will be much more capable.

But when it comes to building worlds in virtual reality, that dreamlike nature might not be such a bad thing, says Snavely. Were used to seeing super-slick and realistic worlds on film and in video games, but theres not quite that level of expectation when it comes to VR. You dont need total photorealism, he says.

Reference: arxiv.org/abs/1707.09405

More on these topics:

Continued here:

AI artist conjures up convincing fake worlds from memories - New Scientist

By and

August 14, 2017, 1:44 PM EDT

Chris Lattner, a legend in the world of Apple software, has joined another rival of the iPhone maker: Alphabet Inc.s Google, where he will work on artificial intelligence.

Lattner announced the news on Twitter on Monday, saying he will start next week. His arrival at Mountain View, California-based Google comes after a brief stint as head of the automated driving program at Tesla Inc., which he left in June. Lattner made a name for himself during a decade-plus career at Apple Inc., where he created the popular programming language Swift.

Exclusive insights on technology around the world.

Get Fully Charged, from Bloomberg Technology.

Lattner said he is joining Google Brain, the search giants research unit. There he will work on a different software language: TensorFlow, Googles system designed to simplify the programming steps for AI, according to a person with knowledge of the matter. Since Google released the software for free last year, it has become a key part of its strategy to spread and make money from AI. Last May, Google introduced a specialized chip set catered for the software, called a TPU, that rents through its cloud-computing service.

A Google spokesman didnt immediately respond to a request for comment.

After leaving Apple in January, Lattner went to Tesla, a recruiting coup for Chief Executive Officer Elon Musk. Lattner left after six months. In the end, Elon and I agreed that he and I did not work well together and that I should leave, so I did, he wrote in an update to his resume.

For more on artificial intelligence, check out the Decrypted podcast:

Go here to read the rest:

Google Hires Former Star Apple Engineer for Its AI Team - Bloomberg

Andrew Keen is the author of three books: Cult of the Amateur, Digital Vertigo and The Internet Is Not The Answer. He produces Futurecast, and is the host of Keen On.

There are few more credible authorities on artificial intelligence (AI) thanHilary Mason the New York-based founder and chief executive of the data science and machine learning consultancyFast Forward Labs.

So, I asked Mason, who is also theData Scientist in Residenceat Accel Partners and theformer Chief Scientist at Bitly, whether todays AI revolution is for real? Or is it, I wondered, just another catch-all phrase used by entrepreneurs and investors to describe the latest Silicon Valley mania?

Mason who sees AI as theumbrella term to describemachine learning andbig data acknowledges that it has become avery trendy area of start-up activity. That said, she says, there has been such rapid technological progress in machine learning over the last five years to make the fieldlegitimately exciting. This progress has been so profound, Mason insists, that it is making AIclose to the heart of our new innovation economy.

But in contrast withthe fearsof prominent technologists like Elon Musk, Mason doesnt worry about the threat to the human species of super intelligent machines. We humans, she says, use machines as tools and the advent of AI doesnt change this.Machines arent rational, she thus argues, implying that there are many more important things for us to worry about than an imminent singularity.

What does concern Mason, however, are questions about the role of women in tech. Thats a question interviewers like myself should be asking men rather than women, she insists. It just createsextra burden for female technologists and thus isnt something that she wants to publicly discuss.

Many thanks to the folks at theGreater Providence Chamber of Commercefor their help in producing this interview.

Continued here:

Why AI is now at the heart of our innovation economy | TechCrunch - TechCrunch