chemical-language-models

#chemical-language-models

ReMedI: Inference-Time Refinement and Self-Distillation for Controllable Molecular Optimization

arXiv cs.LG ↗ · 2026-10-27 Cached

ReMedI 采用基于 Metropolis–Hastings 的推理时搜索策略，让冻结的化学语言模型在多目标药物先导化合物优化中提出并筛选分子编辑，再将搜索轨迹自蒸馏回基础模型，在 C-MuMOInstruct 和 SMDD 基准上以约 7.8× 更少的训练数据达到约 2.8× 于 RL 基线 C-MORAL 的严格成功率。

0 favorites 0 likes

#chemical-language-models

Uncertainty-aware reinforcement learning for chemical language models

arXiv cs.LG ↗ · 2026-06-25 Cached

Proposes two complementary approaches to incorporate predictive uncertainty into reinforcement learning for chemical language models, improving robustness and increasing true hit rate by 0.25 in de novo molecular design.

0 favorites 0 likes

chemical-language-models

ReMedI: Inference-Time Refinement and Self-Distillation for Controllable Molecular Optimization

Uncertainty-aware reinforcement learning for chemical language models

Submit Feedback