Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Friday February 10 2017, @12:51PM   Printer-friendly
from the think-about-it dept.

Google is experimenting to see whether its game-playing AIs will learn to cooperate with each other:

When our robot overlords arrive, will they decide to kill us or cooperate with us? New research from DeepMind, Alphabet Inc.'s London-based artificial intelligence unit, could ultimately shed light on this fundamental question.

They have been investigating the conditions in which reward-optimizing beings, whether human or robot, would choose to cooperate, rather than compete. The answer could have implications for how computer intelligence may eventually be deployed to manage complex systems such as an economy, city traffic flows, or environmental policy.

Joel Leibo, the lead author of a paper DeepMind published online Thursday, said in an e-mail that his team's research indicates that whether agents learn to cooperate or compete depends strongly on the environment in which they operate.

While the research has no immediate real-world application, it would help DeepMind design artificial intelligence agents that can work together in environments with imperfect information. In the future, such work could help such agents navigate a world full of intelligent entities -- both human and machine -- whether in transport networks or stock markets.

DeepMind blog post. Also at The Verge.


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 3, Interesting) by DannyB on Friday February 10 2017, @02:45PM

    by DannyB (5839) Subscriber Badge on Friday February 10 2017, @02:45PM (#465482) Journal

    What if you do something resembling a genetic algorithm at a high level.

    Do a series of competitions between variations of the original AI. The one winning the most times survives the competition of this pair. The survivors of each pair become the candidates for the next generation of testing.

    At the next round of testing, you create new AIs with random tweaks from the most successful of the previous generation. These new AIs and the best of the previous generation all compete to find the new best new AIs. You keep the previous best of the last generation in case the new AIs all do worse than the previous generation.

    --
    People today are educated enough to repeat what they are taught but not to question what they are taught.
    Starting Score:    1  point
    Moderation   +1  
       Interesting=1, Total=1
    Extra 'Interesting' Modifier   0  
    Karma-Bonus Modifier   +1  

    Total Score:   3  
  • (Score: 2) by HiThere on Saturday February 11 2017, @12:21AM

    by HiThere (866) Subscriber Badge on Saturday February 11 2017, @12:21AM (#465640) Journal

    The trick here is how you manage the selection criteria. If the selection criteria reward being cooperative, then that's what you'll get. If they reward being competitive, then *that* is what you'll get. If they reward being "kind" or "caring", then that's what you'll get. But notice just how difficult it is to usefully (i.e. operationally) define those last terms. Cooperative or aggressive are relatively easy.

    --
    Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.