DeepAI AI Chat
Log In Sign Up

Acquisition of Chess Knowledge in AlphaZero

11/17/2021
by   Thomas McGrath, et al.
80

What is learned by sophisticated 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.

READ FULL TEXT

page 17

page 19

page 23

page 37

page 38

page 39

page 40

page 42

02/07/2019

Automating Interpretability: Discovering and Testing Visual Concepts Learned by Neural Networks

Interpretability has become an important topic of research as more machi...
12/31/2022

Mapping Knowledge Representations to Concepts: A Review and New Perspectives

The success of neural networks builds to a large extent on their ability...
07/20/2022

Overlooked factors in concept-based explanations: Dataset choice, concept salience, and human capability

Concept-based interpretability methods aim to explain deep neural networ...
04/19/2023

Learning Hierarchically-Structured Concepts II: Overlapping Concepts, and Networks With Feedback

We continue our study from Lynch and Mallmann-Trenn (Neural Networks, 20...
08/12/2002

Knowledge Representation

This work analyses main features that should be present in knowledge rep...
08/09/2021

Network of scientific concepts: empirical analysis and modeling

Concepts in a certain domain of science are linked via intrinsic connect...
07/31/2021

A Hypothesis for the Aesthetic Appreciation in Neural Networks

This paper proposes a hypothesis for the aesthetic appreciation that aes...