Google DeepMind AI Project Makes Progress
July 25, 2016
For anyone following the development of artificial intelligence, I recommend checking out the article, “How Google Plans to Solve Artificial Intelligence” at MIT Technology Review. The article delves into Google’s DeepMind project, an object of renewed curiosity after its AlphaGo software bested the human world champion of the ancient game Go in March.
This Go victory is significant, because it marks progress beyond the strategy of calculating different moves’ possible outcomes; the game is too complex for that established approach (though such calculations did allow IBM’s DeepBlue to triumph over the world chess champion in 1997). The ability to master Go has some speaking of “intuition” over calculation. Just how do you give software an approximation of human intuition? Writer Tom Simonite tells us:
“Hassabis believes the reinforcement learning approach is the key to getting machine-learning software to do much more complex things than the tricks it performs for us today, such as transcribing our words, or understanding the content of photos. ‘We don’t think just observing is enough for intelligence, you also have to act,’ he says. ‘Ultimately that’s the only way you can really understand the world.’”
“DeepMind’s 3-D environment Labyrinth, built on an open-source clone of the first-person-shooter Quake, is designed to provide the next steps in proving that idea. The company has already used it to challenge agents with a game in which they must explore randomly generated mazes for 60 seconds, winning points for collecting apples or finding an exit…. Future challenges might require more complex planning—for example, learning that keys can be used to open doors. The company will also test software in other ways, and is considering taking on the video game Starcraft and even poker. But posing harder and harder challenges inside Labyrinth will be a major thread of research for some time, says Hassabis. “It should be good for the next couple of years,” he says.”
The article has a video of DeepMind’s virtual labyrinth you can check out, if you’re curious. (It looks very much like an old Windows screen saver some readers may recall.) Simonite tells us that AI firms across the industry are watching this project carefully. He also points to some ways DeepMind is already helping with real-world problems, like developing training software with the U.K.’s National Health Service to help medical personnel recognize commonly missed signs of kidney problems.
See the article for much more about Google’s hopes and plans for DeepMind. Simonite concludes by acknowledging the larger philosophical and ethical concerns around artificial intelligence. We’re told DeepMind has its own “internal ethics board of philosophers, lawyers, and businesspeople.” I think it is no exaggeration to say these folks, whom Google indicates it will name someday soon, could have great influence over the nature of our future technology. Let us hope Google chooses wisely.
Cynthia Murrell, July 25, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
There is a Louisville, Kentucky Hidden Web/Dark Web meet up on July 26, 2016. Information is at this link: http://bit.ly/29tVKpx.
A Peek at the DeepMind Research Process
April 14, 2017
Here we have an example of Alphabet Google’s organizational prowess. Business Insider describes how “DeepMind Organises Its AO Researchers Into ‘Strike Teams’ and ‘Frontiers’.” Writer Sam Shead cites a report by Madhumita Murgia as described in the Financial Times. He writes:
Exactly how DeepMind’s researchers work together has been something of a mystery but the FT story sheds new light on the matter. Researchers at DeepMind are divided into four main groups, including a ‘neuroscience’ group and a ‘frontiers’ group, according to the report. The frontiers group is said to be full of physicists and mathematicians who are tasked with testing some of the most futuristic AI theories. ‘We’ve hired 250 of the world’s best scientists, so obviously they’re here to let their creativity run riot, and we try and create an environment that’s perfect for that,’ DeepMind CEO Demis Hassabis told the FT. […]
DeepMind, which was acquired by Google in 2014 for £400 million, also has a number of ‘strike teams’ that are set up for a limited time period to work on particular tasks. Hassabis explained that this is what DeepMind did with the AlphaGo team, who developed an algorithm that was able to learn how to play Chinese board game Go and defeat the best human player in the world, Lee Se-dol.
Here’s a write-up we did about that significant AlphaGo project, in case you are curious. The creative-riot approach Shead describes is in keeping with Google’s standard philosophy on product development—throw every new idea at the wall and see what sticks. We learn that researchers report on their progress every two months, and team leaders allocate resources based on those reports. Current DeepMind projects include algorithms for healthcare and energy scenarios.
Hassabis launched DeepMind in London in 2010, where offices remain after Google’s 2014 acquisition of the company.
Cynthia Murrell, April 14, 2017
Facebook AI pro Throws Shade at DeepMind Headquarters
November 29, 2016
An AI expert at Facebook criticizes Google’s handling of DeepMind, we learn in Business Insider’s article, “Facebook’s AI Guru Thinks DeepMind is Too Far Away from the ‘Mothership’.” Might Yann LeCun, said guru, be biased? Nah. He simply points out that DeepMind’s London offices are geographically far away from Google’s headquarters in California. Writer Sam Shead, on the other hand, observes that physical distance does not hamper collaboration the way it did before this little thing called the Internet came along.
The article reminds us of rumors that Facebook was eying DeepMind before Google snapped it up. When asked, LeCun declined to confirm or deny that rumor. Shead tells us:
LeCun said: ‘You know, things played out the way they played out. There’s a lot of very good people at DeepMind.’ He added: ‘I think the nature of DeepMind eventually would have been quite a bit different from what it is now if DeepMind had been acquired by a different company than Google.
Google and Facebook are competitors in some areas of their businesses but the companies are also working together to advance the field of AI. ‘It’s very nice to have several companies that work on this space in an open fashion because we build on each other’s ideas,’ said LeCun. ‘So whenever we come up with an idea, very often DeepMind will build on top of it and do something that’s better and vice versa. Sometimes within days or months of each other we work on the same team. They hire half of my students.
Hooray for cooperation. As it happens, London is not an arbitrary location for DeepMind. The enterprise was founded in 2010 by two Oxbridge grads, Demis Hassabis and Mustafa Suleyman, along with UCL professor Shane Legg. Google bought the company in 2014, and has been making the most of their acquisition ever since. For example, Shead reminds us, Google has used the AI to help boost the efficiency of their data-center cooling units by some 40%. A worthy endeavor, indeed.
Cynthia Murrell, November 29, 2016
Sponsored by ArnoldIT.com, publisher of the CyberOSINT monograph
-
- Subscribe to Beyond Search
Feature archive
News archive
-
