Rethinking Entropy Interventions in RLVR: An Entropy Change Perspective

Zhezheng Hao, Hong Wang, Haoyang Liu, Jian Luo, Jiarui Yu, Hande Dong, Qiang Lin, Can Wang, Jiawei Chen

Annual Meeting of the Association for Computational Linguistics

We propose a quantitative analysis framework for entropy change. Building on this, the effect of entropy interventions can be unified and elucidated through token-level analysis. Our findings point out a fundamental limitation of existing methods: they attempt to control the entropy indirectly. By only affecting related factors, such as the advantage signal and generation probability, their effectiveness is inherently limited and could potentially fail.

Can Data-Driven Dynamics Reveal Hidden Physics? There Is A Need for Interpretable Neural Operators

Wenhan Gao*, Jian Luo*, Fang Wan, Ruichen Xu, Xiang Liu, Haipeng Xing, Yi Liu (* equal contribution)

We provide a way to explain the prediction-making process of neural operators and show that neural operator can learn hidden physical patterns from data. However, this explanation method is limited to specific situations, highlighting the urgent need for generalizable explanation methods. Next, we show that a simple dual-space multi-scale model can achieve SOTA performance and we believe that dual-space multi-spatio-scale models hold significant potential to learn complex physics and require further investigation. Lastly, we discuss the critical need for principled frameworks to incorporate known physics into neural operators, enabling better generalization and uncovering more hidden physical phenomena.

Neural Krylov Iteration for Accelerating Linear System Solving

Jian Luo, Jie Wang, Hong Wang, Huanshuo Dong, Zijie Geng, Hanzhu Chen, Yufei Kuang

Advances in Neural Information Processing Systems

We propose a novel method, namely Neural Krylov Iteration (NeurKItt), for accelerating linear system solving. To enhance the subspace prediction accuracy, we utilize QR decomposition for the neural operator outputs and introduce a novel projection loss function for training. NeurKItt accelerates the solving of linear systems across various settings and datasets, achieving up to a 5.5× speedup in computation time and a 16.1× speedup in the number of iterations.