Monthly Archives: February 2017

Slide: 1 / of 1. Caption: Getty Images

Artificial intelligence is not one thing, but many, spanning several schools of thought. In his book The Master Algorithm, Pedro Domingos calls them the tribes of AI.

As the University of Washington computer scientist explains, each tribe fashions what would seem to be very different technology. Evolutionists, for example, believe they can build AI by recreating natural selection in the digital realm. Symbolists spend their time coding specific knowledge into machines, one rule at a time.

Right now, the connectionists get all the press. They nurtured the rise of deep neural networks, the pattern recognition systems reinventing the likes of Google, Facebook, and Microsoft. But whatever the press says, the other tribes will play their own role in the rise of AI.

Take Ben Vigoda, the CEO and founder of Gamalon. Hesa Bayesian, part of the tribe that believes in creating AI through the scientific method. Rather than building neural networks that analyze data and reach conclusions on their own, he and his team useprobabilistic programming, a technique in which they start with their own hypotheses and then use data to refine them. His startup, backed by Darpa, emerged from stealth mode this morning.

Gamalons tech can translate from one language to another, and the company isdevelopingtools that businesses can use to extract meaning from raw streams of text. Vigoda claims his particular breed of probabilistic programming can produce AI that learns more quickly than neural networks, using much smaller amounts of data. You can be very careful about what you teach it, he says, and can edit what youve taught it.

As others point out, an approach along these lines is essential to the rise of machines capable of truly thinking like humans. Neural networks require enormous amounts of carefully labelled data, and this isnt always available. Vigoda even goes so far as to say that his techniques will replace neural networks completely, in all applications. That is very, very clear, he says.

But just as deep learning isnt the only way to artificial intelligence, neither is probabilistic programming. Or Gaussian processes. Or evolutionary computation. Or reinforcement learning.

Sometimes, the AI tribesbadmouth each other. Sometimes, they play up their technology at the expense of the others. But the reality is that AI will risefrom many technologies working together. Despite the competition, everyone is working toward the same goal.

Probabilistic programming lets researchers build machine learning algorithms more like coders build computer programs. But the real power of the technique lies inits ability to deal with uncertainty. This can allow AI to learn from less data, but it can also helpresearchers understand why an AI reaches particular decisionsand more easily tweak the AI if they dont agree with those decisions. True AI will need all that, whether it powers a chatbot trying to carry on a human-like conversation or an autonomous car trying to avoid an accident.

But neural networks have proven their worth with, among other things, image and speech recognition, and theyre not necessarily in competition with techniques like probabilistic programming. In fact, Google researchers are building systems that combine the two. Their strengths complement one another. Deep neural networks and probabilistic models are closely related, says David Blei, a Columbia University computer scientist and an advisor to Gamalon who has worked with Google research on these types of mixed models. Theres a lot of probabilistic modeling happening inside neural networks.

Inevitably, the best AI will combine several technologies. Take AlphaGo, the breakthrough system built by Googles DeepMind lab. It combined neural networks with reinforcement learning and other techniques. Blei, for one, doesnt see a world oftribes. It doesnt exist for me, he says. He sees a world in which everyone is reaching for the same master algorithm.

Read the rest here:

AI's Factions Get Feisty. But Really, They're All on the Same Team - WIRED

Forbes

IoT And AI: Improving Customer Satisfaction Forbes Truethe Internet of Things (IoT) and artificial intelligence (AI) hold huge promise in helping us better engage and satisfy our customers. But that promise still depends heavily on our ability to process and act on the data we're gathering in a way ...

Continue reading here:

IoT And AI: Improving Customer Satisfaction - Forbes

Many experts in the field firmly believe 2017 will be a breakout year for both artificial intelligence and robotics, since the two often go together. Spoiler alert: it's all good.

AI Makes Robots Smarter

Robots use an increasing number of sensing modalities including taste, smell, sonar, IR, haptic feedback, tactile sensors, and range of motion sensors. They are also becoming better at picking up on facial expressions and gestures, so their interactions with humans become more natural, said Kevin Curran, IEEE senior member and professor of cyber security at Ulster University.

"Basically, AI is crucial for all their learning and adaptive behavior so they can adapt existing capabilities to cope with environmental changes. AI is key to helping them learn new tasks on the fly by sequencing existing behaviors," he said.

Karsten Schmidt, head of technology at the Innovation Center Silicon Valley for SAP Labs echoed this sentiment. "In 2017, we will see AI gain greater acceptance and momentum as humans come to increasingly rely, trust and depend more on AI-driven decisions and question them less. This will happen as a direct result of improved AI learning due to more usage and a broader user base, and as the quality and usefulness of AI software in turn improves," he said.

Meet Your AI Co-Worker

Many people fear losing their jobs to robots, but more than likely you will have a robot for a co-worker. Then again, if you've been in the workforce long enough, you've probably already had a robot for a co-worker, just in human form.

"In 2017, we are seeing a growing emergence of robots designed to operate alongside people in everyday human environments. Autonomous service robots that assist workers in warehouses, deliver supplies in hospitals, and maintain inventory of items in grocery stores are emerging onto the market," said Sonia Chernova, assistant professor at Georgia Tech College of Computing.

These systems need humans because one thing robotics researchers are still struggling with is robotic arms. There's no substitute for the human arm to pick things up and manipulate objects. "[Robot arms] have of course been used successfully for decades in manufacturing, but current techniques work reliably only in controlled factory environments, and are not yet robust enough for the real world," said Chernova.

This could lead to the rise of "AI Supervisors," said Tomer Naveh, CTO of Adgorithms, an AI-based digital marketing platform. Robots already have taken on many labor-intensive, manual (read: boring) tasks we do in our everyday life but robots will get smarter, and need AI to do it, he said.

"AI systems will get better at communicating their decisions and reasoning to their operators, and those operators will respond with new rules, business logic, and feedback that make it more and more useful in practice over time. As a result we will see people shifting from doing tasks by themselves, to supervising AI software on how to do it for them," he said.

That's actually a disturbing thought.

Changing Retail

AI and robotics will slowly move into another area where human error is common: retail. To some degree there is already automation in optical scanners and retail tracking used by stores to manage inventory, but it will be considerably improved.

The retail industry, for example, has been unable to address the problem of non-scanned items at checkout, which accounts for 30% of retailers annual losses. They only discover the loss in inventory well after the fact.

"AI is stepping in to address issues of this caliber across industries, and as a result, its often gathering just as much data as its processing. This resulting data is becoming a secondary benefit to businesses that use AI. AI Apps created to detect these non-scans are now also providing retailers with information about their origins, whether theyre fraudulent or accidental, and how customers and cashiers are gaming the system," said Alan OHerlihy, CEO of Everseen, developer of AI products for point of sale systems.

And as consumers have positive experiences with drone deliveries, public opinion may go a long way towards opening up regulations for further drone use, said Jake Rheude, director of business development for Red Stag Fulfillment, an eCommerce fulfillment provider.

"Consumers are already fully on board with the concept of drone delivery. According to The Walker Sands Future of Retail 2016 Study, 79% of US consumers said they would be 'very likely' or 'somewhat likely' to request drone delivery if their package could be delivered within an hour. And 73% of respondents said that they would pay up to $10 for a drone delivery. This is an unprecedented level of acceptance for new technology with so little real word experience from consumers," he said.

AI in Your Home

Another prediction made by umpteen science fiction movies usually with an alarmist tone is that AI will come into the home in a big way. It already has if you have an iPhone, with Siri, or use Windows 10 and Cortana. Gradually it will move into other devices, the experts predict.

"Alexa, Cortana and Siri are great, but they still lack the sophistication and accuracy to be relied upon as a utility. In 2017, advances in natural language processing and natural language generation will transform what digital assistants understand and how they analyze and respond with legitimately useful information. The era of just opening a related Wikipedia page are over," said Matt Gould, AI expert and co-founder of Arria NLG, which develops technology that translates data into language.

To make these devices work optimally, they need to develop an emotional quotient, or an EQ, predicts Dr. Rana el Kaliouby, CEO and co-founder, Affectiva, which develops facial recognition software. "We expect to see Emotion AI really come to the fore this year, and once AI systems develop social skills and rapport, AI interfaces will be more engaging and sticky, and less frustrating for their users, driving even wider adoption of the technology," she said.

She predicts that in the future, all of our devices will be equipped with a chip that can adapt our experiences to our emotions in real time, by reading facial expressions, analyzing tone of voice and possessing built-in emotion awareness. "The ability of technology to adapt to our mood and preferences could enhance experiences ranging from driving a car to ordering a pizza," she said.

And this should mean less typing, said Scott Webb, president of Avionos. "Physical interaction with hand-to-keyboard commands will give way to more organic input methods like voice and physical response as we move forward," he said.

Better Security

It's been said before but is worth repeating that AI will improve security because, like in so many other cases, security AI won't be prone to human failings of boredom, fatigue, illness and disinterest that often causes a security lapse. It will also have much faster reaction times and much better recognition of unusual patterns.

"Machine learning and the models generated through processes around machine learning are helping enterprises analyze massive amounts of data and identify trends, anomalies, and things not detectable through standard modeling. Machine learning algorithms are helping security researchers dynamically identify threats, airlines improve maintenance and reliability of their aircraft, and provide the back bone for self-driving cars to analyze data in real-time to make decisions," said David Dufour, senior director of engineering at antimalware vendor WebRoot.

That immediacy is needed with catching data breaches, as well. The average time to discover a network attacker is about five months, giving attackers plenty of time to achieve their goals, said Peter Nguyen, director of technical services at LightCyber, which does behavior based security software.

"Finding signs of an attacker is difficult and demands the use of AI. Instead of trying to encounter, identify and block threats by their known characteristics, the way to find an active attacker is through their operational activities. Using machine learning, its possible to learn the good behavior of all users and devices and then find anomalies. Then, AI can be focused to find those anomalies that are truly indicative of an active attack," he said.

More:

AI and Robotics Trends: Experts Predict - Datamation

A consumers credit score used to be a commonly understood number the time-honored FICO score that banks all used in their underwriting. But banks increasingly are relying on dozens of scores that reflect a variety of data sources, analytics and use of artificial intelligence technology.

The use of AI offers lenders the ability to get a precise look into someones creditworthiness and score those previously deemed unscorable.

But such scoring techniques also bring uncertainty: What it will take to convince regulators that AI-based credit scores are not a black box? How do you get a system trained to look at the interactions of many variables, to produce one clear reason for declining credit? Data scientists at credit bureaus and banks are working to find answers to questions like these.

The benefits of AI-powered credit scores

There are two main reasons to use artificial intelligence to derive a credit score. One is to assess creditworthiness more precisely. The other is to be able to consider people who might not have been able to get a credit score in the past, or who may have been too hastily rejected by a traditional logistic regression-based score. In other words, a method that looks at certain data points from consumers credit history to calculate the odds that they will repay.

[Digital identity is broken, and fixes are urgently needed. Learn how large financial service and healthcare companies are tackling the issue to enhance customer experience, to stake out positions in their business ecosystems, and to manage risk on our Feb. 23 web seminar. Click here for details.]

Machine learning can take a more nuanced look at consumer behavior.

A neural network more closely mimics the way humans think and reason, whereas linear models are more dogmatic youre imposing structure on data as opposed to letting the data talk to you, said Eric VonDohlen, chief analytics officer at the online lender Elevate. The more complex reasoning of artificial intelligence can find things in the data that wouldnt be apparent otherwise.

And instead of considering one variable at a time, an artificial intelligence engine can look at interactions between multiple variables.

Its harder for the workhorse, logistic regression, to do that, said Dr. Stephen Coggeshall, chief analytics and science officer at ID Analytics. You have to do a lot of data preprocessing using expert knowledge to even attempt to find those nonlinear interactions.

Consumers with several chargeoffs in their histories would most likely be considered high-risk borrowers by most traditional models. But an AI engine might perceive mitigating variables; though the consumers might have skipped payments on three debts in the past 24 months, they have paid on time consistently for the past year and have successfully obtained new lines of credit.

It looks like that bad performance or bad history is in your past, VonDohlen said. That would be a simple example of how an AI world might help cast data in a more positive and more accurate light.

AI-based credit scoring models let Elevate make sharper predictions of credit risk, approve the right people and offer better pricing to people who deserve it, VonDohlen said.

Elevate is deploying its new, AI-based models gradually, starting with 1% of potential borrowers, testing the results, and gradually applying them to more people.

Credit bureaus are starting to adopt AI in their credit scores, too.

Equifax calls the machine-learning software that it uses in credit scores NeuroDecision Technology.

Technologies like Hadoop, which allow massive amounts of data to be stored and analyzed quickly, are making AI-based credit scores possible, said Peter Maynard, senior vice president of global analytics at Equifax.

Before, if you gave me a million observations, it would take a week to sort through it, Maynard said.

ID Analytics uses what it calls convolutional neural nets, a flavor of deep learning, in its fraud and credit scores, Coggeshall said. For its Credit Optics Full Spectrum credit score, AI engines look at consumer payment data from wireless, utility and marketplace loan providers, to score consumers who have thin or no credit bureau files, including young people and new credit seekers.

FICO also offers a score called XD thats based on telephone and utility bill payments and property records. It gives high marks to people who have faithfully paid their phone, oil and gas bills and who have not moved around too much.

Experian is taking a more cautious approach. It uses traditional logistic regression methods for its credit scores, but in its labs it experiments with machine learning.

When the technology seems to make a significant difference in performance, the company will provide credit scores based on machine learning, said Eric Haller, executive vice president of Experians Global DataLabs. For now, he sees machine learning giving only a nominal lift in results.

The opportunity is not building the next VantageScore, because believe it or not, those scores work really well, he said.

To let clients experiment with machine learning, Experian offers an analytical sandbox with its credit data loaded into it.

They can load their own data in and well sync it up with historical credit archive data, and weve overlaid it with a set of machine-learning tools, Haller said. Most large financial institutions are using the sandbox today, he said.

TransUnion, the other major credit rating agency, did not respond to requests for an interview.

Now for the confusing part

The cons of AI-enhanced credit scores include the risk that the full underwriting process will be hidden from consumers and that the practice would raise transparency questions among regulators.

Last month, the Consumer Financial Protection Bureau imposed $23 million in fines to TransUnion and Equifax, noting they claim banks use their scores to determine creditworthiness, when that isnt always the case.

In their advertising, TransUnion and Equifax falsely represented that the credit scores they marketed and provided to consumers were the same scores lenders typically use to make credit decisions, the CFPB said in a press release announcing the fines. In fact, the scores sold by TransUnion and Equifax were not typically used by lenders to make those decisions.

Some say the regulator was misguided.

Both scores being sold by TransUnion and Equifax, VantageScore and the Equifax RiskScore, are real credit scores that are Equal Credit Opportunity Act compliant, are commercially available to lenders and are, in fact, used by lenders, said credit expert John Ulzheimer.

Equifax says it ran all of its AI-based scoring technology past the OCC, Fed and CFPB and got a positive response. ID Analytics said it worked closely with lawyers, compliance officers and regulators to assure the technology complied with various lending rules.

Another challenge to using artificial intelligence, specifically neural networks, in credit scores and models, is that its harder to provide the needed reason code to borrowers the explanation of why they were denied credit.

Concerns about the reason code are the main reason many businesses dont use nonlinear machine-learning models for credit scores yet.

A lot of the confusion and heartburn is around, How do you boil an extremely data-rich learning process into a marginal rationale for declining a loan? VonDohlen said.

However, neural networks, which are essentially designed to think like a brain, can also be used to help find the one variable that represents the greatest risk.

Its almost never the case that you would decline someone for a rats nest of variable relationships, VonDohlen said. The reasons for credit denial, he said, "are almost always very clear.

Equifax has developed a proprietary algorithm that can generate reason codes for consumers, Maynard said.

Experian is also working on techniques that would make AI credit-based score decisions more explainable and auditor friendly.

Were not operating under any assumption that a black box credit scoring model would even work or be accepted in the market, Haller said. We are 100% focused on how do we bridge the gap such that we can bring better performance to models, but still maintain the same integrity, where they can be explained to the OCC and our clients are comfortable with understanding how the models are working and the results theyre getting.

But in the end, consumers wont be confused, according to Ulzheimer.

Regardless of how many scoring systems are being used, they are all based on three credit reports, he said. If you've got three great credit reports, then every single scoring system being used is going to yield a high score.

Penny Crosman is Editor at Large at American Banker.

More:

Is AI making credit scores better, or more confusing? - American Banker

SHA-1 End Times Have Arrived

January 17, 2017 , 11:00 am

January 3, 2017 , 4:28 pm

December 29, 2016 , 11:30 am

February 14, 2017 , 3:44 pm

February 13, 2017 , 11:00 am

February 10, 2017 , 11:45 am

February 8, 2017 , 8:21 am

February 3, 2017 , 11:20 am

February 2, 2017 , 2:57 pm

January 30, 2017 , 4:48 pm

January 26, 2017 , 11:16 am

January 23, 2017 , 8:52 am

January 20, 2017 , 11:50 am

January 13, 2017 , 10:00 am

January 11, 2017 , 4:40 pm

January 10, 2017 , 11:28 am

January 6, 2017 , 12:00 pm

January 4, 2017 , 2:01 pm

December 22, 2016 , 6:00 am

December 19, 2016 , 1:42 pm

December 13, 2016 , 3:27 pm

December 12, 2016 , 1:47 pm

December 9, 2016 , 11:00 am

December 8, 2016 , 9:15 am

December 6, 2016 , 11:24 am

December 5, 2016 , 2:10 pm

December 1, 2016 , 12:00 pm

November 30, 2016 , 12:44 pm

November 28, 2016 , 3:30 pm

November 8, 2016 , 2:57 pm

November 1, 2016 , 5:50 pm

October 29, 2016 , 6:00 am

October 27, 2016 , 4:27 pm

October 25, 2016 , 3:00 pm

October 22, 2016 , 6:00 am

October 21, 2016 , 10:01 am

October 20, 2016 , 7:00 am

October 18, 2016 , 4:58 pm

October 14, 2016 , 9:00 am

October 5, 2016 , 8:51 am

October 3, 2016 , 5:00 am

September 26, 2016 , 10:45 am

September 22, 2016 , 3:47 pm

September 22, 2016 , 12:31 pm

September 20, 2016 , 2:41 pm

September 15, 2016 , 11:15 am

September 13, 2016 , 9:14 am

September 9, 2016 , 2:06 pm

September 8, 2016 , 3:43 pm

September 2, 2016 , 9:00 am

September 1, 2016 , 1:08 pm

August 29, 2016 , 9:58 am

August 24, 2016 , 5:53 pm

August 24, 2016 , 8:00 am

August 17, 2016 , 4:06 pm

August 17, 2016 , 12:58 pm

August 8, 2016 , 1:40 pm

August 4, 2016 , 3:26 pm

August 4, 2016 , 10:00 am

August 3, 2016 , 10:00 am

August 2, 2016 , 9:00 am

July 29, 2016 , 10:45 am

July 26, 2016 , 9:30 am

July 25, 2016 , 3:51 pm

July 21, 2016 , 1:18 pm

July 20, 2016 , 9:21 am

July 15, 2016 , 11:00 am

July 14, 2016 , 1:05 pm

July 12, 2016 , 11:40 am

June 30, 2016 , 11:48 am

June 28, 2016 , 10:00 am

May 31, 2016 , 5:44 pm

May 31, 2016 , 1:37 pm

March 10, 2016 , 10:23 am

December 23, 2016 , 5:19 pm

December 27, 2016 , 1:22 pm

February 13, 2017 , 9:00 am

December 28, 2016 , 4:00 am

December 28, 2016 , 9:00 am

See the article here:

Cryptographers Dismiss AI, Quantum Computing Threats - Threatpost

CNET

Watch Drive.ai's self-driving car handle California city streets on a ... TechCrunch Autonomous vehicle system startup Drive.ai is ready to show off its technology in use for the first time, with a video showing the company's test vehicle.. Watch Drive.ai's car drive itself on a dark, rainy night - RoadshowCNET Watch Drive.ai's Deep Learning Car Drive Itself AroundInverse A rainy night is no trouble for this self-driving carThe Verge Silicon Valley Business Journal -YouTube -Medium all 22 news articles »

See the article here:

Watch Drive.ai's self-driving car handle California city streets on a ... - TechCrunch

Human translators will face off against artificially intelligent (AI) machine translators next week in Seoul, South Korea.

The competition, sponsored by Sejong Cyber University and the International Interpretation and Translation Association (IITA) will pit human translators against Google Translate and Naver Papago.

Google Translate and Naver Papago are two of the most popular English-Korean AI translation services.Both use Neural Machine Translation (NMT), which replaces traditional word-for-word systems with full-sentence translation, helping to improve syntax and flow in machine translations.

Neural Machine Translation also creates machines that are capable of self-improvement through deep learning and big data analysis.

The challenge for competitors will be to translate two randomly selected news articles from English to Korean, and two from Korean to English. Each competitor will have 30 minutes per article to complete the translation, which will be judged for accuracy. Judges have been selected from professors at Sejong University and professional translators with the IITA.

Sejong University will live stream the event. The expectation is that the artificial intelligence programs will win for speed, but that the human translators will win for accuracy. However, in the research paper published by Google with the initial announcement of Neural Machine Translation, humanversus-AI tests showed that the AI program was able to translate random sentences with near-human accuracy.

IITA Secretary-General Kang Dae-young said, We hope to confirm that humans and machines have different strengths and weaknesses and highlight that human professionals will still have their roles in translation and interpretation of the future.

In one study comparing Google Translate to Naver Papago in English-Korean translations, it was found that Naver was better at translating slang, while Google was better at long sentences. Both performed well in translating short sentences accurately.

With advances in deep learning, human versus AI competitions are growing in popularity. Just last month, a poker competition at Carnegie Mellon University pitted humans against an AI program called Libratus. The 20-day competition ended with the machine winning by a resounding $1.76 million in chips over human competitors. A different program, DeepStack, recently became the first to beat humans at no-limit poker.

Read the rest here:

Humans engage AI in translation competition - The Stack