arXiv cs.AI· Kaihui Cheng, Zhiqiang Cai, Wenkai Xiang, Zhihang Hu, Siyu Zhu, Tzuhsiung Yang, Yuan Qi·· 9 天前AI 评分25
在生成空间中学习隐式偏置以加速蛋白质动力学模拟
Learning Implicit Bias in Generative Spaces for Accelerating Protein Dynamics Emulation
AI 导读
研究者提出一种在预训练蛋白质动力学生成模拟器的生成空间中注入隐式、依赖历史的偏置方法，通过历史感知评分估计器引导逆向采样远离已生成结构，并用基于评分的精修步骤将漂移样本重新投影回数据流形。
正文 · 原文
Abstract:Generative emulators of protein dynamics produce plausible trajectories at a fraction of the cost of molecular dynamics, but they inherit their training distribution and tend to revisit known states rather than reach rare ones under long-horizon extrapolation. Inspired by classical enhanced sampling, we introduce an implicit, history-dependent bias in the generative space of a pretrained emulator. Specifically, a history-aware score estimator augments the frozen emulator with a distance-weighted bias that steers reverse-time sampling away from previously generated structures, regularized by an environment-support term. To preserve structural validity at long horizons, a score-based refinement step re-projects drifted samples onto the data manifold using the frozen emulator. Our experiments demonstrate that the method (i) raises diversity by $35\%$ on DynamicPDB-80; (ii) on $12$ zero-shot Fast-Folding proteins, the learned bias alone reaches the unbiased emulator's coverage up to ${\sim}15\times$ faster, and pairing it with refinement reaches the coverage up to ${\sim}37\times$ faster while covering ${\sim}3\times$ as many low-energy states. Code will be released soon.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2606.01833 [cs.LG] |
| (or arXiv:2606.01833v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2606.01833 arXiv-issued DOI via DataCite |
Submission history
From: Kaihui Cheng [view email]
[v1]
Mon, 1 Jun 2026 07:46:28 UTC (2,223 KB)
[v2]
Thu, 8 Oct 2026 07:34:26 UTC (2,228 KB)
来源：arXiv cs.AI · arxiv.org