Representation Matters for Mastering Chess: Improved Feature Representation in AlphaZero Outperforms Switching to Transformers
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2023
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866914916996218880 |
|---|---|
| author | Czech, Johannes Blüml, Jannis Kersting, Kristian Steingrimsson, Hedinn |
| author_facet | Czech, Johannes Blüml, Jannis Kersting, Kristian Steingrimsson, Hedinn |
| contents | 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 improvements, our experimental results using the Integrated Gradient technique confirm the effectiveness of the newly introduced features. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2304_14918 |
| institution | arXiv |
| publishDate | 2023 |
| record_format | arxiv |
| spellingShingle | Representation Matters for Mastering Chess: Improved Feature Representation in AlphaZero Outperforms Switching to Transformers Czech, Johannes Blüml, Jannis Kersting, Kristian Steingrimsson, Hedinn Artificial Intelligence 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 improvements, our experimental results using the Integrated Gradient technique confirm the effectiveness of the newly introduced features. |
| title | Representation Matters for Mastering Chess: Improved Feature Representation in AlphaZero Outperforms Switching to Transformers |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2304.14918 |