Proceedings of the 20th Conference on Computer Science and Intelligence Systems (FedCSIS)

DOI: http://dx.doi.org/10.15439/2025F3227

Citation: Proceedings of the 20th Conference on Computer Science and Intelligence Systems (FedCSIS), M. Bolanowski, M. Ganzha, L. Maciaszek, M. Paprzycki, D. Ślęzak (eds). ACSIS, Vol. 43, pages 813–818 (2025)

Full text

Abstract. The goal of FedCSIS 2025 Challenge is to build a model to predict the difficulty (measured as Lichess rating) of given chess puzzles. To address this task, we propose a three-stage joint visual--statistical framework for predicting Glicko-based difficulty ratings. In the first stage, a convolutional model based on MobileNetV2 integrates FEN-rendered board images with structured features, including engine-predicted success probabilities, move count, and piece counts, to generate baseline predictions. The second stage employs LightGBM to perform residual refinement, explicitly learning the residual errors of the baseline predictions to correct systematic biases, particularly for extreme difficulty levels. Finally, a domain-informed refinement adjusts the outputs toward interpretable difficulty estimates derived from failure probability distributions and rating-bucket inflection points. Our model ranked 9th in the challenge. Experimental results show that residual refinement and domain-informed adjustment significantly reduce mean squared error compared to the baseline visual--statistical model.

References

1. Lichess.org, “Puzzle Rating System,” Available: https://lichess.org/training/about, accessed Jan. 2025.
2. J. Zyśko, M. Ślęzak, D. Ślęzak, and M. Świechowski, “FedCSIS 2025 knowledgepit.ai Competition: Predicting Chess Puzzle Difficulty Part 2 & A Step Toward Uncertainty Contests,” in Proc. 20th Conf. Comput. Sci. Intell. Syst. (FedCSIS), vol. 43, Polish Inf. Process. Soc., 2025. http://dx.doi.org/10.15439/2025F5937.
3. J. Zyśko, M. Świechowski, S. Stawicki, K. Jagieła, A. Janusz and D. Śl˛ ezak, "IEEE Big Data Cup 2024 Report: Predicting Chess Puzzle Difficulty at KnowledgePit.ai," 2024 IEEE International Conference on Big Data (BigData), Washington, DC, USA, 2024, pp. 8423-8429, https://dx.doi.org/10.1109/BigData62323.2024.10825289.
4. D. Ruta, M. Liu and L. Cen, "Moves Based Prediction of Chess Puzzle Difficulty with Convolutional Neural Networks," 2024 IEEE International Conference on Big Data (BigData), Washington, DC, USA, 2024, pp. 8390-8395, https://dx.doi.org/10.1109/BigData62323.2024.10825595.
5. S. Miłosz and P. Kapusta, “Transformer-based Modeling of Chess Puzzle Difficulty from Sequential Data,” arXiv preprint https://arxiv.org/abs/2403.01234, 2024.
6. M. Omori and P. Tadepalli, “Joint Puzzle and Player Rating Estimation with CNN-LSTM Models,” in Proc. IEEE BigData Cup, 2024.
7. T. Woodruff, O. Filatov, and M. Cognetta, “Bread Emoji Team: Hybrid Prediction of Chess Puzzle Difficulty Using Maia/Leela and Tree-Based Regressors,” in Proc. IEEE BigData Cup, 2024.
8. A. Schütt, T. Huber, and E. André, “Human Problem-Solving Inspired Neural Networks for Chess Puzzle Difficulty Prediction,” in Proc. IEEE BigData Cup, 2024.
9. M. Sandler, A. Howard, M. Zhu, A. Zhmoginov, and L.-C. Chen, “MobileNetV2: Inverted Residuals and Linear Bottlenecks,” arXiv preprint https://arxiv.org/abs/1801.04381, 2019. [Online]. Available: https://arxiv.org/abs/1801.04381
10. G. Ke et al., “LightGBM: A Highly Efficient Gradient Boosting Decision Tree,” in Advances in Neural Information Processing Systems (NeurIPS), 2017, pp. 3146–3154.