What next for the computer Go program?


Readers!
 
For many reasons, mostly political but partly ethical, I do not use Google, Facebook, Twitter. They practice corrupt business policies, while targeting conservative websites for censoring, facts repeatedly confirmed by news stories and by my sense that Facebook has taken action to prevent my readers from recommending Behind the Black to their friends.
 
Thus, I must have your direct support to keep this webpage alive. Not only does the money pay the bills, it gives me the freedom to speak honestly about science and culture, instead of being forced to write it as others demand.

 

Please consider donating by giving either a one-time contribution or a regular subscription, as outlined in the tip jar below.


 

Regular readers can support Behind The Black with a contribution via paypal:

Or with a subscription with regular donations from your Paypal or credit card account:


If Paypal doesn't work for you, you can support Behind The Black directly by sending your donation by check, payable to Robert Zimmerman, to
 
Behind The Black
c/o Robert Zimmerman
P.O.Box 1262
Cortaro, AZ 85652

 

You can also support me by buying one of my books, as noted in the boxes interspersed throughout the webpage. And if you buy the books through the ebookit links, I get a larger cut and I get it sooner.

Link here.

The software uses neural networks to learn from experience. For example, to train for its Go match the computer program studied 30 million Go board positions from human games, then played itself again and again to improve its skills.

DeepMind’s founder and chief executive Demis Hassabis mentioned the possibility of training a version of AlphaGo using self-play alone, omitting the knowledge from human-expert games, at a conference last month. The firm created a program that learned to play less complex arcade games in this manner in 2015. Without a head start, AlphaGo would probably take much longer to learn, says Bengio — and might never beat the best human. But it’s an important step, he says, because humans learn with such little guidance.

DeepMind, based in London, also plans to venture beyond games. In February the company founded DeepMind Health and launched a collaboration with the UK National Health Service: its algorithms could eventually be applied to clinical data to improve diagnoses or treatment plans. Such applications pose different challenges from games, says Oren Etzioni, chief executive of the non-profit Allen Institute for Artificial Intelligence in Seattle, Washington. “The universal thing about games is that you can collect an arbitrary amount of data,” he says — and that the program is constantly getting feedback on what’s a good or bad move by playing many games. But, in the messy real world, data — on rare diseases, say — might be scarcer, and even with common diseases, labelling the consequences of a decision as ‘good’ or ‘bad’ may not be straightforward.

Hassabis has said that DeepMind’s algorithms could give smartphone personal assistants a deeper understanding of users’ requests. And AI researchers see parallels between human dialogue and games: “Each person is making a play, and we have a sequence of turns, and each of us has an objective,” says Bengio. But they also caution that language and human interaction involve a lot more uncertainty.

Share

2 comments

  • Wayne

    Mr. Z— “very cool”
    This makes a lot more sense to me now.

    (Not my area at all & very leery in general of the science-media attributing mentalistic & anthropomorphic attributes to machines.)

    ((very interested in how concepts of “emergent-order” would manifest themselves within Programming & AI designs.))

  • Dick Eagleson

    For an early, but still impressive, demonstration of “emergent order” I would recommend looking into the Game of Life which dates back to the 60’s. More recent work that also reflects “emergent order” are so-called “genetic algorithms” that work toward an optimized end-point by generating random solutions and competing them against one another, with the winners getting to breed more of the next generation of candidate solutions than the losers.

Leave a Reply

Your email address will not be published. Required fields are marked *