AlphaZero, general reinforcement learning algorithm, Google DeepMind, London, United Kingdom

[Airicist]

Developer - Google DeepMind (https://pr.ai/showthread.php?4751)

chessprogramming.org/AlphaZero (https://www.chessprogramming.org/AlphaZero)

AlphaZero (https://en.wikipedia.org/wiki/AlphaZero) on Wikipedia

AlphaGo (https://pr.ai/showthread.php?13908), computer Go program

MuZero (https://pr.ai/showthread.php?t=22741), gaming program

[Airicist]

Article "Mastering Chess and Shogi by Self-Play with a General Reinforcement Learning Algorithm (https://arxiv.org/abs/1712.01815)"

by David Silver (https://pr.ai/showthread.php?t=21795), Thomas Hubert, Julian Schrittwieser, Ioannis Antonoglou, Matthew Lai, Arthur Guez, Marc Lanctot, Laurent Sifre, Dharshan Kumaran, Thore Graepel, Timothy Lillicrap, Karen Simonyan, Demis Hassabis
December 5, 2017

[Airicist]

AlphaZero beats AlphaGo Zero, Stockfish, and Elmo (http://talkchess.com/forum/viewtopic.php?topic_view=threads&p=741340&t=65911)

[Airicist]

Article "The future is here – AlphaZero learns chess (https://en.chessbase.com/post/the-future-is-here-alphazero-learns-chess)"

by Albert Silver
December 6, 2017

[Airicist]

Article "Google's AlphaZero Destroys Stockfish In 100-Game Match (https://www.chess.com/news/view/google-s-alphazero-destroys-stockfish-in-100-game-match)"

by Mike Klein
December 6, 2017

[Airicist]

Article "AlphaZero AI beats champion chess program after teaching itself in four hours (https://www.theguardian.com/technology/2017/dec/07/alphazero-google-deepmind-ai-beats-champion-program-teaching-itself-to-play-four-hours)"
Google’s artificial intelligence sibling DeepMind repurposes Go-playing AI to conquer chess and shogi without aid of human knowledge

by Samuel Gibbs
December 7, 2017

[Airicist]

Article "Alpha Zero’s “Alien” Chess Shows the Power, and the Peculiarity, of AI (https://www.technologyreview.com/s/609736/alpha-zeros-alien-chess-shows-the-power-and-the-peculiarity-of-ai)"
The latest advance from DeepMind behaves in a very surprising way. Expect other AI systems to be just as odd.

by Will Knight
December 8, 2017

[Airicist]

"How to build your own AlphaZero AI using Python and Keras (https://medium.com/applied-data-science/how-to-build-your-own-alphazero-ai-using-python-and-keras-7f664945c188)"

by David Foster
January 26, 2018

[Airicist]

https://youtu.be/7L2sUGcOgh0

AlphaZero: Shedding new light on the grand games of chess, shogi and Go (https://deepmind.com/blog/alphazero-shedding-new-light-grand-games-chess-shogi-and-go)

Published on Dec 6, 2018


DeepMind's AlphaZero is the successor of AlphaGo, the first computer program to beat a world champion at the ancient game of Go. It taught itself from scratch how to master the games of chess, shogi and Go, beating a world-champion program in each case and discovering new and creative playing strategies that hint at the potential of these systems to tackle other complex problems.

[Airicist]

"AlphaZero: Shedding new light on the grand games of chess, shogi and Go (https://deepmind.com/blog/alphazero-shedding-new-light-grand-games-chess-shogi-and-go)"

[Airicist]

Book "Game Changer: AlphaZero's Groundbreaking Chess Strategies and the Promise of AI (https://www.amazon.com/Game-Changer-AlphaZeros-Groundbreaking-Strategies/dp/9056918184)"

by Matthew Sadler, Natasha Regan, Garry Kasparov
January 20, 2019

[Airicist]

https://youtu.be/nPexHaFL1uo

AlphaZero's attacking chess

Premiered Dec 6, 2018


Google's DeepMind has just released a new academic paper on AlphaZero -- the general purpose artificial intelligence system that mastered chess through self-play and went on to defeat the world champion of chess engines, Stockfish. In this video chess International Master Anna Rudolf takes a look at a never-before-seen game from a match played in January 2018, and discusses how the playing style and attacking chess of AlphaZero compare to computers and humans.

The game I selected is part of the 20-game collection me and other chess broadcasters received before the release of the PGN.

[Airicist]

https://youtu.be/7L2sUGcOgh0

AlphaZero: Shedding new light on the grand games of chess, shogi and Go (https://deepmind.com/blog/alphazero-shedding-new-light-grand-games-chess-shogi-and-go)

Published on Dec 6, 2018


DeepMind's AlphaZero is the successor of AlphaGo, the first computer program to beat a world champion at the ancient game of Go. It taught itself from scratch how to master the games of chess, shogi and Go, beating a world-champion program in each case and discovering new and creative playing strategies that hint at the potential of these systems to tackle other complex problems.

Article "AlphaZero Crushes Stockfish In New 1,000-Game Match (https://www.chess.com/news/view/updated-alphazero-crushes-stockfish-in-new-1-000-game-match)"

by pete
December 6, 2018

[Airicist]

https://youtu.be/BazNQEeqNhU

Visiting the DeepMind Headquarters: My AlphaZero Challenge

Published on Dec 20, 2018


DeepMind's AlphaZero shook the chess world in December 2017 by mastering the game from scratch: after only 4 hours of self-play the AI system was capable of beating the strongest chess computer, Stockfish, in a 100-game match. A year later DeepMind released a full evaluation of AlphaZero in the journal Science -- occasion on which chess International Master Anna Rudolf visits the DeepMind headquarters to figure out more about the "AlphaZero effect". The survey she conducted at the London Chess Classic, a super tournament held a floor above DeepMind's offices at Google, presents the opinion of renowned figures of the chess community on the AI system. The twist? Each interviewee was limited by Anna's challenge: Describe AlphaZero in one sentence.

[Airicist]

https://youtu.be/1gWpFuQlBsg

AlphaZero: DeepMind’s AI works smarter, not harder

Published on Feb 26, 2019

[Airicist]

Article "AlphaZero beat humans at Chess and StarCraft, now it’s working with quantum computers (https://thenextweb.com/artificial-intelligence/2020/01/16/alphazero-beat-humans-at-chess-and-starcraft-now-its-working-with-quantum-computers)"

by Tristan Greene
January 16, 2020

[Airicist]

https://youtu.be/hNw7whJqsJ8

Chess Grandmasters on Google Deepmind AlphaZero || Artificial Intelligence in Chess

Jul 5, 2020


Chess grandmasters share their opinion on Google Deepmind AlphaZero in chess.

The chess masters Maxime Vachier-Lagrave (MVL), Jan Gustafsson, Alexei Shirov, Sergei Movsesian and Andreas Heimann talk about the AlphaZero vs Stockfish match and discuss the impact of AI in chess.

In the end of 2017 AlphaZero has beaten the chess engine Stockfish. This was the first time that a chess artificial intelligence based on Reinforcement Learning could beat the strongest chess engine.

0:00 - AlphaZero vs Stockfish
5:38 - Garri Kasparov vs Deep Blue 1997

[Airicist]

Article "AI Ruined Chess. Now, It’s Making the Game Beautiful Again (https://www.wired.com/story/ai-ruined-chess-now-making-game-beautiful)"
A former world champion teams up with the makers of AlphaZero to test variants on the age-old game that can jolt players into creative patterns.

by Tom Simonite
September 9, 2020

[Airicist]

https://youtu.be/O1b0cbgpRBw

Assessing Game Balance with AlphaZero: Exploring Alternative Rule Sets in Chess (Paper Explained)

Sep 13, 2020


Chess is a very old game and both its rules and theory have evolved over thousands of years in the collective effort of millions of humans. Therefore, it is almost impossible to predict the effect of even minor changes to the game rules, because this collective process cannot be easily replicated. This paper proposes to use AlphaZero's ability to achieve superhuman performance in board games within one day of training to assess the effect of a series of small, but consequential rule changes. It analyzes the resulting strategies and sets the stage for broader applications of reinforcement learning to study rule-based systems.

OUTLINE:
0:00 - Intro & Overview
2:30 - Alternate Chess Rules
4:20 - Using AlphaZero to assess rule change outcomes
6:00 - How AlphaZero works
16:40 - Alternate Chess Rules continued
18:50 - Game outcome distributions
31:45 - e4 and Nf3 in classic vs no-castling chess
36:40 - Conclusions & comments

Paper: https://arxiv.org/abs/2009.04374

Abstract:
It is non-trivial to design engaging and balanced sets of game rules. Modern chess has evolved over centuries, but without a similar recourse to history, the consequences of rule changes to game dynamics are difficult to predict. AlphaZero provides an alternative in silico means of game balance assessment. It is a system that can learn near-optimal strategies for any rule set from scratch, without any human supervision, by continually learning from its own experience. In this study we use AlphaZero to creatively explore and design new chess variants. There is growing interest in chess variants like Fischer Random Chess, because of classical chess's voluminous opening theory, the high percentage of draws in professional play, and the non-negligible number of games that end while both players are still in their home preparation. We compare nine other variants that involve atomic changes to the rules of chess. The changes allow for novel strategic and tactical patterns to emerge, while keeping the games close to the original. By learning near-optimal strategies for each variant with AlphaZero, we determine what games between strong human players might look like if these variants were adopted. Qualitatively, several variants are very dynamic. An analytic comparison show that pieces are valued differently between variants, and that some variants are more decisive than classical chess. Our findings demonstrate the rich possibilities that lie beyond the rules of modern chess.

Authors: Nenad Tomašev, Ulrich Paquet, Demis Hassabis, Vladimir Kramnik

[Airicist]

Article "DeepMind's AI is helping to re-write the rules of the chess (https://www.zdnet.com/article/deepminds-ai-is-helping-to-re-write-the-rules-of-the-chess)"
DeepMind's researchers are letting AlphaZero play with different rules to find out how to improve the game.

by Daphne Leprince-Ringuet
September 14, 2020

[Airicist]

"AlphaZero, a novel Reinforcement Learning Algorithm, in JavaScript (https://towardsdatascience.com/alphazero-a-novel-reinforcement-learning-algorithm-deployed-in-javascript-56018503ad18)"
Learn about and implement AlphaZero, entirely in JavaScript!

by Carlos Aguayo
November 8, 2020

[Airicist2]

Article "DeepMind makes bet on AI system that can play poker, chess, Go, and more (https://venturebeat.com/2021/12/08/deepmind-makes-bet-on-ai-system-that-can-play-poker-chess-go-and-more)"

by Kyle Wiggers (https://www.linkedin.com/in/kyle-lee-wiggers)
December 8, 2021

[Airicist2]

"Discovering novel algorithms with AlphaTensor (https://www.deepmind.com/blog/discovering-novel-algorithms-with-alphatensor)"

October 5, 2022

[Airicist2]

"Acquisition of chess knowledge in AlphaZero (https://www.pnas.org/doi/10.1073/pnas.2206625119)"

by Thomas McGrath, Andrei Kapishnikov, Nenad Tomašev, and Vladimir Kramnik
November 14, 2022