Acquisition of Chess Knowledge in AlphaZero

• URL: http://arxiv.org/abs/2111.09259v1
• Date: Wed, 17 Nov 2021 17:46:19 GMT
• Title: Acquisition of Chess Knowledge in AlphaZero
• Authors: Thomas McGrath and Andrei Kapishnikov and Nenad Toma\v{s}ev and Adam Pearce and Demis Hassabis and Been Kim and Ulrich Paquet and Vladimir Kramnik
• Abstract summary: We show that human knowledge is acquired by the AlphaZero neural network as it trains on the game of chess. By probing for a broad range of human chess concepts we show when and where these concepts are represented in the AlphaZero network. We also provide a behavioural analysis focusing on opening play, including qualitative analysis from chess Grandmaster Vladimir Kramnik.
• Score: 14.41428465712717
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: What is being learned by superhuman neural network agents such as AlphaZero? This question is of both scientific and practical interest. If the representations of strong neural networks bear no resemblance to human concepts, our ability to understand faithful explanations of their decisions will be restricted, ultimately limiting what we can achieve with neural network interpretability. In this work we provide evidence that human knowledge is acquired by the AlphaZero neural network as it trains on the game of chess. By probing for a broad range of human chess concepts we show when and where these concepts are represented in the AlphaZero network. We also provide a behavioural analysis focusing on opening play, including qualitative analysis from chess Grandmaster Vladimir Kramnik. Finally, we carry out a preliminary investigation looking at the low-level details of AlphaZero's representations, and make the resulting behavioural and representational analyses available online.

Related papers

• Evidence of Learned Look-Ahead in a Chess-Playing Neural Network [11.746104876318606]
  We present evidence of learned look-ahead in the policy network of Chess Leela Zero. We find that Leela internally represents future optimal moves and that these representations are crucial for its final output in certain board states. These findings are an existence proof of learned look-ahead in neural networks and might be a step towards a better understanding of their capabilities.
  arXiv  Detail & Related papers  (2024-06-02T21:57:32Z)
• Simple and Effective Transfer Learning for Neuro-Symbolic Integration [50.592338727912946]
  A potential solution to this issue is Neuro-Symbolic Integration (NeSy), where neural approaches are combined with symbolic reasoning. Most of these methods exploit a neural network to map perceptions to symbols and a logical reasoner to predict the output of the downstream task. They suffer from several issues, including slow convergence, learning difficulties with complex perception tasks, and convergence to local minima. This paper proposes a simple yet effective method to ameliorate these problems.
  arXiv  Detail & Related papers  (2024-02-21T15:51:01Z)
• Bridging the Human-AI Knowledge Gap: Concept Discovery and Transfer in AlphaZero [10.569213003371654]
  We show how to extract new chess concepts from AlphaZero, an AI system that mastered the game of chess via self-play without human supervision. In a human study, we show that these concepts are learnable by top human experts, as four top chess grandmasters show improvements in solving the presented concept prototype positions. This marks an important first milestone in advancing the frontier of human knowledge by leveraging AI.
  arXiv  Detail & Related papers  (2023-10-25T06:49:26Z)
• Evaluation Beyond Task Performance: Analyzing Concepts in AlphaZero in Hex [39.001544338346655]
  We investigate AlphaZero's internal representations in the game of Hex using two evaluation techniques from natural language processing (NLP): model probing and behavioral tests. We find that concepts related to short-term end-game planning are best encoded in the final layers of the model, whereas concepts related to long-term planning are encoded in the middle layers of the model.
  arXiv  Detail & Related papers  (2022-11-26T21:59:11Z)
• Neural Networks for Chess [2.055949720959582]
  AlphaZero, Leela Chess Zero and Stockfish NNUE revolutionized Computer Chess. This book gives a complete introduction into the technical inner workings of such engines.
  arXiv  Detail & Related papers  (2022-09-03T22:17:16Z)
• An AlphaZero-Inspired Approach to Solving Search Problems [63.24965775030674]
  We adapt the methods and techniques used in AlphaZero for solving search problems. We describe possible representations in terms of easy-instance solvers and self-reductions. We also describe a version of Monte Carlo tree search adapted for search problems.
  arXiv  Detail & Related papers  (2022-07-02T23:39:45Z)
• Searching for the Essence of Adversarial Perturbations [73.96215665913797]
  We show that adversarial perturbations contain human-recognizable information, which is the key conspirator responsible for a neural network's erroneous prediction. This concept of human-recognizable information allows us to explain key features related to adversarial perturbations.
  arXiv  Detail & Related papers  (2022-05-30T18:04:57Z)
• Neuro-Symbolic Learning of Answer Set Programs from Raw Data [54.56905063752427]
  Neuro-Symbolic AI aims to combine interpretability of symbolic techniques with the ability of deep learning to learn from raw data. We introduce Neuro-Symbolic Inductive Learner (NSIL), an approach that trains a general neural network to extract latent concepts from raw data. NSIL learns expressive knowledge, solves computationally complex problems, and achieves state-of-the-art performance in terms of accuracy and data efficiency.
  arXiv  Detail & Related papers  (2022-05-25T12:41:59Z)
• Towards Human-Understandable Visual Explanations:Imperceptible High-frequency Cues Can Better Be Removed [46.36600006968488]
  We argue that the capabilities of humans, constrained by the Human Visual System (HVS) and psychophysics, need to be taken into account. We conduct a case study regarding the classification of real vs. fake face images, where many of the distinguishing features picked up by standard neural networks turn out not to be perceptible to humans.
  arXiv  Detail & Related papers  (2021-04-16T08:11:30Z)
• Explainability in Deep Reinforcement Learning [68.8204255655161]
  We review recent works in the direction to attain Explainable Reinforcement Learning (XRL) In critical situations where it is essential to justify and explain the agent's behaviour, better explainability and interpretability of RL models could help gain scientific insight on the inner workings of what is still considered a black box.
  arXiv  Detail & Related papers  (2020-08-15T10:11:42Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.