Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Sunday June 04 2017, @02:11PM   Printer-friendly
from the what-killed-the-cat? dept.

http://www.sciencemag.org/news/2017/05/scientists-imbue-robots-curiosity

Over the years, scientists have worked on algorithms for curiosity, but copying human inquisitiveness has been tricky. For example, most methods aren't capable of assessing artificial agents' gaps in knowledge to predict what will be interesting before they see it. (Humans can sometimes judge how interesting a book will be by its cover.)

Todd Hester, a computer scientist currently at Google DeepMind in London hoped to do better. "I was looking for ways to make computers learn more intelligently, and explore as a human would," he says. "Don't explore everything, and don't explore randomly, but try to do something a little smarter."

So Hester and Peter Stone, a computer scientist at the University of Texas in Austin, developed a new algorithm, Targeted Exploration with Variance-And-Novelty-Intrinsic-Rewards (TEXPLORE-VENIR), that relies on a technique called reinforcement learning. In reinforcement learning, a program tries something, and if the move brings it closer to some ultimate goal, such as the end of a maze, it receives a small reward and is more likely to try the maneuver again in the future. DeepMind has used reinforcement learning to allow programs to master Atari games and the board game Go through random experimentation. But TEXPLORE-VENIR, like other curiosity algorithms, also sets an internal goal for which the program rewards itself for comprehending something new, even if the knowledge doesn't get it closer to the ultimate goal.


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This discussion has been archived.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: 2) by captain normal on Sunday June 04 2017, @04:16PM (3 children)

    by captain normal (2205) on Sunday June 04 2017, @04:16PM (#520251)

    How does one "reward" a machine? Feed it more amps? Play with it's dongle?

    --
    When life isn't going right, go left.
    Starting Score:    1  point
    Karma-Bonus Modifier   +1  

    Total Score:   2  
  • (Score: 2) by VLM on Sunday June 04 2017, @05:33PM

    by VLM (445) Subscriber Badge on Sunday June 04 2017, @05:33PM (#520283)

    It seems self evident when dealing with neural networks.

    My guess for non-neural network systems is you drip money sex food and power into the programmers who are editing the source code until the output is indistinguishable from the programmer getting laid for the AI's response being so perfect.

    That's kinda how the cheating market works in education although thats a little more explicit and never seems to result in strong AI. Which implies that model of software development might not work for AI.

    Neural networks are not a magic panacea either as extensive experience indicates that it takes millions of neural networks that can't even code fizzbuzz to find the couple that are halfway decent programmers.

  • (Score: 0) by Anonymous Coward on Sunday June 04 2017, @05:37PM (1 child)

    by Anonymous Coward on Sunday June 04 2017, @05:37PM (#520288)

    > How does one "reward" a machine? Feed it more amps? Play with it's dongle?

    It is really subjective, but we mostly like to watch sexy pinball machines or start global thermonuclear war.

    • (Score: 0) by Anonymous Coward on Monday June 05 2017, @06:57AM

      by Anonymous Coward on Monday June 05 2017, @06:57AM (#520601)

      I didn't realize you were all Republicans. Maybe it's time to let robot AIs vote in elections, I mean openly.