Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Thursday October 19, @02:39PM   Printer-friendly
from the Zeroing-in-on-AI dept.

Google DeepMind researchers have made their old AlphaGo program obsolete:

The old AlphaGo relied on a computationally intensive Monte Carlo tree search to play through Go scenarios. The nodes and branches created a much larger tree than AlphaGo practically needed to play. A combination of reinforcement learning and human-supervised learning was used to build "value" and "policy" neural networks that used the search tree to execute gameplay strategies. The software learned from 30 million moves played in human-on-human games, and benefited from various bodges and tricks to learn to win. For instance, it was trained from master-level human players, rather than picking it up from scratch.

AlphaGo Zero did start from scratch with no experts guiding it. And it is much more efficient: it only uses a single computer and four of Google's custom TPU1 chips to play matches, compared to AlphaGo's several machines and 48 TPUs. Since Zero didn't rely on human gameplay, and a smaller number of matches, its Monte Carlo tree search is smaller. The self-play algorithm also combined both the value and policy neural networks into one, and was trained on 64 GPUs and 19 CPUs over a few days by playing nearly five million games against itself. In comparison, AlphaGo needed months of training and used 1,920 CPUs and 280 GPUs to beat Lee Sedol.

Though self-play AlphaGo Zero even discovered for itself, without human intervention, classic moves in the theory of Go, such as fuseki opening tactics, and what's called life and death. More details can be found in Nature, or from the paper directly here. Stanford computer science academic Bharath Ramsundar has a summary of the more technical points, here.

Go is an abstract strategy board game for two players, in which the aim is to surround more territory than the opponent.

Previously: Google's New TPUs are Now Much Faster -- will be Made Available to Researchers
Google's AlphaGo Wins Again and Retires From Competition


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This discussion has been archived.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 DannyB on Thursday October 19, @06:59PM

    by DannyB (5839) on Thursday October 19, @06:59PM (#584752)

    it would still have been somewhat more impressive, and useful, if it had discovered things we didn't already know.

    If it did already, would we be able to recognize it?

    The reason we know it discovered various game playing tragedies that we have names for, is because we recognize those particular game tragedies. How would we recognize that it discovered a hence unknown and unnamed game tragedy?

    . . . as well as novel strategies that provide new insights into the oldest of games.

    Novel strategies ... Sounds like they won't be very useful. Have they previously been discovered by man and dismissed due to their level of novelty?

    Did you catch the part about providing new insights?

    Maybe these novel tragedies the machine discovered, which provide new insights, are indeed useful and represent an advancement in human knowledge.

    Starting Score:    1  point
    Karma-Bonus Modifier   +1  

    Total Score:   2