Reality is merely an illusion, albeit a very persistent one.

Benchmark Alpha Zero

1 May 2018
Download as .zip Download as .tar .gz View on GitHub

by Saleem Ahmed

tags: reinforcement-learning actor-critic deep-learning chess p-np

One experiment is simply to give AlphaZero an old-fashioned examination on test positions for which the perfect answers are known. These could even be generated in a controlled fashion from chess endgames with 7 or fewer pieces on the board, for which perfect play was tabulated by Victor Zakharov and Vladimir Makhnichev using the Lomonosov supercomputer of Moscow State University. Truth in those tables is often incredibly deep—in some positions the win takes over 500 moves, many of which no current chess program (not equipped with the tables) let alone human player would find. Or one can set checkmate-in-{N} problems that have stumped programs to varying degrees.

Apha Zero Benchmark

Overview

One experiment is simply to give AlphaZero an old-fashioned examination on test positions for which the perfect answers are known.

These could even be generated in a controlled fashion from chess endgames with 7 or fewer pieces on the board, for which perfect play was tabulated by Victor Zakharov and Vladimir Makhnichev using the Lomonosov supercomputer of Moscow State University.

Truth in those tables is often incredibly deep—in some positions the win takes over 500 moves, many of which no current chess program (not equipped with the tables) let alone human player would find.

Or one can set checkmate-in-{N} problems that have stumped programs to varying degrees.

The Candidate Game :

A game playable with chess pawns or checkers on a board with 5 rows and {n} columns, where perhaps {n} can be set to achieve the small-enough/large-enough balance. Dr. Regan conceived this 35 years ago at Oxford when {n=4} seemed right for computers of the day. The starting position is:

Starting Position

Goal :

Make your opponent run out of legal moves.

Rules :

Launching the Web App for playing the game/agent

We suggest that you begin by typing:

  1. cd ./zero_truth/
  2. npm start

    npm start

    Starts the development server.

    npm run build

    Bundles the app into static files for production.

    npm test

    Starts the test runner.

    npm run eject

    Removes this tool and copies build dependencies, configuration files and scripts into the app directory. If you do this, you can’t go back!

Referrences :

  1. Truth From Zero

Happy Hacking ! :boom: :trollface: :boom:

back