Categories
GeekWire

IBM Watson AI XPRIZE offers $5 million

Image: IBM Watson
IBM’s Watson AI software is best-known for winning at “Jeopardy!” in 2011. (Credit: IBM)

The latest multimillion-dollar tech challenge – known as the IBM Watson AI XPRIZE – will be aimed at encouraging collaboration between humans and artificial intelligence software to solve the world’s big problems. But it’s not yet clear who will sign up for the $5 million competition, in part because IBM’s Watson program is already one of the contestants in a much bigger, multibillion-dollar AI race.

The Feb. 17 announcement, made at the TED2016 conference in Vancouver, B.C., adds artificial intelligence to an XPRIZE list that also spotlights ocean discovery, moon exploration, carbon recycling, medical diagnostic devices, educational software and much, much more.

“Our hope is that the teams will show how we can apply AI to the world’s great challenges,” Stephanie Wander, who’s on the prize development team for California-based XPRIZE, told GeekWire. “That would be the cat’s meow.”

A lot of the details surrounding the competition still have to be worked out. The complete rules and guidelines are to be made available in mid-May, just before IBM’s World of Watson conference. Teams can already pre-register.

Get the full story on GeekWire.

Categories
GeekWire

Odds are shifting in man-vs.-machine Go match

Image: Go match
European Go champion Fan Hui holds his head in frustration during a match against Google DeepMind’s AlphaGo program. (Credit: Google via YouTube)

WASHINGTON, D.C. – Both sides in next month’s big $1 million AI-vs.-human Go match say they’re confident they’ll prevail. But Google DeepMind’s AlphaGo program has a secret weapon: It’s expanding its knowledge of the game exponentially during the buildup to the five-game match against top-ranked player Lee Sedol in Seoul, South Korea.

Last month, researchers at Google DeepMind shook up the Go world with news that its artificial intelligence program bested a European champion, Fan Hui, without being given any advantage to start with. The research, published in Nature, lays out a potentially more powerful approach to AI that combines deep learning with reinforcement learning.

Next month’s match against Sedol could be as big for fans of Go (and followers of AI research) as IBM’s Deep Blue victory over chess champion Garry Kasparov was in 1997. The match will be streamed live from Seoul via You Tube from March 9 to 15.

“This really is our Deep Blue moment,” Demis Hassabis, Google DeepMind’s president of engineering, said on Feb. 13 at the American Association for the Advancement of Science’s annual meeting in Washington.

Get the full story on GeekWire.

Categories
GeekWire

AI experts say robots could spark unemployment

Image: Google self-driving car
Google is testing subcompact self-driving cars. (Photo via Google)

WASHINGTON, D.C. – The robot revolution may put half of us humans out of a job by 2045 – and if that happens, what are the politicians going to do about it?

“This issue of automation and employment, which is going to be one of the biggest policy issues for the next 25 years, if not longer, and now we’re in a presidential election year … this issue is just nowhere on the radar screen,” Rice University computer scientist Moshe Vardi said Feb. 13 at the American Association for the Advancement of Science’s annual meeting in Washington.

Vardi and other experts on artificial intelligence sketched out a scary picture of what the next couple of decades could bring as machines become smarter, more powerful and more prevalent. It’s a picture that’s developing quickly, thanks to the rise of machine vision and machine learning.

Bart Selman, a computer science professor at Cornell University, said he would not have been as concerned about AI’s downside five years ago. Since then, however, engineers have brought about dramatic improvements in the ability of software systems to see, hear and understand their environment.

Get the full story on GeekWire.

Categories
GeekWire

A.I. software masters the game of Go

Image: Go computer
Google is set for the ultimate human-vs.-machine Go match. (Credit: Nature / Google DeepMind)

Mark another milestone in the rise of the machines: An artificial intelligence program pioneered by Google DeepMind has learned how to play the game of Go well enough to beat a human champion decisively in a fair match.

That’s a quantum leap for artificial intelligence: Go is looked upon as the “holy grail of AI research,” said Demis Hassabis, the senior author of a research paper on the project published today by the journal Nature.

The game seems simple enough, involving the placement of alternating black and white stones on a 19-by-19 grid. The object is merely to avoid having your stones hemmed in on four sides by your opponent’s stones. But Go, which originated in China thousands of years ago, is considered the world’s most complex game. “It has 10170 possible board positions, which is greater than the number of atoms in the universe,” Hassabis noted.

That means a computer program can’t best humans with the same kind of approach used for checkers and and chess. The programs for those games combine brute-force searches through the possible moves with a weighted evaluation of patterns in moves. But researchers at Google DeepMind say their software, known as AlphaGo, takes a different approach.

Get the full story on GeekWire.

Categories
GeekWire

A.I. pioneer Marvin Minsky dies at age of 88

Image: Marvin Minsky
MIT Professor Emeritus Marvin Minsky has died at 88. (Credit: Louis Fabian Bachrach via MIT)

Marvin Minsky, the computer scientist who helped blaze the trail for virtual smartphone assistants and other manifestations of artificial intelligence, died of a cerebral hemorrhage on Sunday at the age of 88.

Word of Minsky’s passing came today from his family as well as from the Massachusetts Institute of Technology, where Minsky worked on the foundations of A.I. since the 1950s.

As the co-founder of MIT’s Artificial Intelligence Laboratory, now known as theComputer Science and Artificial Intelligence Laboratory, Minsky contributed to the decades-long drive to make machines more humanlike – through the development of robotic hands and neural networks, as well through his musings on the philosophical underpinnings of intelligence.

Get the full story on GeekWire.

Categories
GeekWire

Experts find quicker way to teach a computer

Image: Human vs. machine
This illustration gives a sense of how characters from alphabets around the world were replicated through human vs. machine learning. (Credit: Danqing Wang)

Researchers say they’ve developed an algorithm that can teach a new concept to a computer using just one example, rather than the thousands of examples that are traditionally required for machine learning.

The algorithm takes advantage of a probabilistic approach the researchers call “Bayesian Program Learning,” or BPL. Essentially, the computer generates its own additional examples, and then determines which ones fit the pattern best.

The researchers behind BPL say they’re trying to reproduce the way humans catch on to a new task after seeing it done once – whether it’s a child recognizing a horse, or a mechanic replacing a head gasket.

“The gap between machine learning and human learning capacities remains vast,” said MIT’s Joshua Tenenbaum, one of the authors of a research paper published today in the journal Science. “We want to close that gap, and that’s the long-term goal.”

Get the full story on GeekWire.

Categories
GeekWire

Let Google’s A.I. bot answer your emails

Image: Encoder
Google’s Smart Reply feature analyzes incoming email with an encoder. (Google image)

Weary of spambots, robo-calls and Twitter bots? Google is coming out with an artificial-intelligence tool that’s on your side for a change: Smart Reply, a feature that’s built into its Inbox app for Android and iOS.

Smart Reply is designed to take the thumbwork out of replying to email on a mobile device.

“I get a lot of email, and I often peek at it on the go with my phone. But replying to email on mobile is a real pain, even for short replies,” Greg Corrado, a senior research scientist at Google, writes on the company’s research blog. “What if there were a system that could automatically determine if an email was answerable with a short reply, and compose a few suitable responses that I could edit or send with just a tap?”

Corrado explains at length how Google’s engineers developed a deep neural network that analyzes incoming email and suggests short responses based on context.

Get the full story on GeekWire.