Skip to main content Skip to main navigation

Publikation

Representation Matters for Mastering Chess: Improved Feature Representation in AlphaZero Outperforms Switching to Transformers

Johannes Czech; Jannis Blüml; Kristian Kersting; Hedinn Steingrimsson
In: Ulle Endriss; Francisco S. Melo; Kerstin Bach; Alberto José Bugarín Diz; Jose Maria Alonso-Moral; Senén Barro; Fredrik Heintz (Hrsg.). ECAI 2024 - 27th European Conference on Artificial Intelligence, 19-24 October 2024, Santiago de Compostela, Spain - Including 13th Conference on Prestigious Applications of Intelligent Systems (PAIS 2024). AAAI Conference on Artificial Intelligence (AAAI), Pages 2378-2385, Frontiers in Artificial Intelligence and Applications, Vol. 392, IOS Press, 2024.

Zusammenfassung

Representation Matters for Mastering Chess: Improved Feature Representation in AlphaZero Outperforms Switching to Transformers Johannes Czech1 Jannis Bl¨uml1,2 Kristian Kersting1,2,3,4 Hedinn Steingrimsson5,6 1 Artificial Intelligence and Machine Learning Lab, TU Darmstadt, Germany 2 Hessian Center for Artificial Intelligence (hessian.AI), Darmstadt, Germany 3 Centre for Cognitive Science, TU Darmstadt, Germany 4 German Research Center for Artificial Intelligence (DFKI), Darmstadt, Germany 5 Department of Electrical and Computer Engineering, Rice University, Houston, TX, USA 6 The Steingrimsson Foundation (safesystem2.org), Houston, TX, USA E-mail: {johannes.czech, blueml, kersting}@cs.tu-darmstadt.de, hedinn.steingrimsson@rice.edu Abstract While transformers have gained recognition as a versatile tool for artificial intelligence (AI), an unexplored challenge arises in the context of chess — a classical AI benchmark. Here, incorporating Vision Transformers (ViTs) into AlphaZero is insufficient for chess mastery, mainly due to ViTs’ computational limitations. The attempt to optimize their efficiency by combining MobileNet and NextViT outperformed AlphaZero by about 30 Elo. However, we propose a practical improvement that involves a simple change in the input representation and value loss functions. As a result, we achieve a significant performance boost of up to 180 Elo points beyond what is currently achievable with AlphaZero in chess. In addition to these improve- ments, our experimental results using the Integrated Gradient technique confirm the effectiveness of the newly introduced features.

Weitere Links