Otter: Time-Aware, History-Conditioned Human Chess AI Released
Summary:ArXiv has released Otter, a lightweight human chess AI model with 15.3 million parameters. By modeling chess as a time-aware and sequential process, Otter significantly improves the accuracy of human move prediction. It combines a move history encoder and a clock pressure modulation module, trained on 117 million Lichess rapid games. Otter outperforms the previous state-of-the-art model, Maia 2, across multiple Elo brackets. This release demonstrates the potential for efficient AI performance im
Otter: Time-Aware, History-Conditioned Human Chess AI
Key Technical Highlights
- Time-Aware Modeling: Otter introduces a time-aware mechanism, treating chess as a sequential process rather than isolated positions. This approach aligns more closely with human decision-making logic.
- History-Conditioning: The model uses a move history encoder, conditioning predictions on the last 20 moves to capture opening preferences, positional drift, and in-game behavioral tendencies.
- Clock Pressure Modulation: A clock pressure module adjusts predictions based on remaining time, allowing the model to adapt to varying time pressures.
- Efficient Training and Lightweight Design: Otter contains only 15.3 million parameters and was trained on a single T4 GPU for 30 days, significantly reducing computational resource requirements.
Performance
- Prediction Accuracy: Otter achieves 55.23% top-1 and 90.95% top-5 move prediction accuracy on Lichess rapid game data, surpassing the previous state-of-the-art model, Maia 2.
- Elo Bracket Performance: Across 11 Elo brackets (<1100 to >=2000), Otter performs best in the 1900-1999 bracket with an accuracy of 57.38%.
Industry Impact and Developer Recommendations
- New Breakthrough in AI Chess: Otter demonstrates the potential of time-aware and history-conditioning modeling to enhance AI chess player performance, providing a new direction for AI chess research.
- Improved Resource Efficiency: The lightweight design and efficient training method make Otter highly practical in resource-constrained environments.
- Developer Recommendations: Developers are encouraged to explore the application of time-aware modeling and history-conditioning methods in other sequence prediction tasks.
Conclusion
The release of Otter marks an important milestone in the field of AI chess. Through innovative time-aware and history-conditioning methods, Otter achieves significant improvements in prediction accuracy and resource efficiency, offering new insights into AI model design and application.
— END —Tags: #Otter #Human Chess AI #Time-Aware Modeling #Lightweight Model
Editorial & Fact-Checking Note: This article is compiled from primary research, official release documentation, and source papers by the ZICQ Newsroom pipeline with automated entity verification and human editorial review. If you notice any technical inaccuracy, please submit a correction via our corrections policy or email our editorial desk directly.
Community Comments