arXiv cs.AI· Jia Yi Zhang, Shivanshu Shekhar, Allison McLevige, Tong Zhang·· 12 小时前AI 评分31
ReMedI：用推理时精炼与自蒸馏实现可控分子优化
ReMedI: Inference-Time Refinement and Self-Distillation for Controllable Molecular Optimization
AI 导读
ReMedI 将分子语言模型作为推理时对齐的基础提议器，用 Metropolis-Hastings 搜索策略引导多目标分子编辑，在 C-MuMOInstruct 和 SMDD 基准上于三个骨干模型上显著提升多目标优化成功率。其搜索轨迹可蒸馏回基础模型，训练所需分子先导物比 C-MORAL 少约 7.8 倍，严格成功率覆盖约为 C-MORAL 的 2.8 倍，并包含分布外任务。
正文 · 原文
Abstract:In drug development, lead optimization must improve competing properties, such as increasing permeability and reducing toxicity, while limiting scaffold changes that could compromise the lead's potency. RL-based post-training, such as C-MORAL, addresses this problem by fine-tuning molecular language models to directly optimize molecular leads for objectives of interest, which relies on expensive task-specific datasets and faces challenges in achieving multi-objective optimization (MOO). We argue a pre-trained model may already contain useful molecular edits that zero-shot generation cannot efficiently reach. Motivated by the limited search capabilities of zero-shot generation, we develop an efficient Metropolis-Hastings search strategy to better explore the proposal distribution. We propose ReMedI, which uses the molecular language model as a base proposer for inference-time alignment. Given a lead and a set of multi-objective tasks, the model proposes edits conditioned on the current molecule, while Metropolis--Hastings refinement guides sequential edits under an objective for property improvement, preservation, and lead similarity. On both C-MuMOInstruct and SMDD benchmarks, refinement improves frozen base models across three backbones, reaching significantly higher multi-objective optimization success rates compared to baselines. These search traces can then distilled into the base model which improves zero-shot generation and subsequent search using approximately 7.8 times less training molecular leads than that of C-MORAL. ReMedI achieves approximately 2.8 times the strict-success coverage of RL baseline C-MORAL, including on out-of-distribution tasks, with matched lead-relative similarity.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.32103 [cs.LG] |
| (or arXiv:2610.32103v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.32103 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jia Yi Zhang [view email]
[v1]
Fri, 23 Oct 2026 03:07:35 UTC (4,818 KB)
来源：arXiv cs.AI · arxiv.org