arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2602.00663cs.AIcs.LGq-bio.BM
收藏

SEISMO:通过轨迹感知的LLM代理提高分子优化的样本效率

SEISMO: Explanation-Aware, Trajectory-Conditioned LLM Agents for Sample-Efficient Molecular Optimisation

Technical University of Munich, Germany(慕尼黑技术大学,德国);TUM School of Computation, Information(TUM计算、信息学院);Technology, Department of Mathematics(技术学院,数学系);Helmholtz Munich – German Research Center for Environmental Health (GmbH), Institute of Structural Biology, Molecular Targets(海德堡慕尼黑德国环境健康研究所以及结构生物学研究所,分子靶点);Therapeutics Center, 85764 Neuherberg, Germany(治疗中心,德国新赫尔伯格85764);Machine Learning Research(机器学习研究)

Fabian P. Krüger, Andrea Hunklinger, Adrian Wolny, Tim J. Adler, Igor Tetko, Santiago David Villalba

AI总结:

SEISMO通过轨迹感知的LLM代理实现高效的分子优化,利用结构化反馈提升样本效率。

摘要:

优化分子结构以实现期望的性质是化学科学中的核心瓶颈,尤其是在制药行业,它在新药发现中起着基础作用。由于分子性质评估通常依赖于昂贵且速率受限的预言机,如实验测试,因此分子优化必须具有高度的样本效率。为了解决这一问题,我们引入了SEISMO,一个能够严格在线进行推理时间分子优化的LLM代理,在每次预言机调用后更新,而不需要基于种群或批量学习。SEISMO将每个提案条件于完整的优化轨迹,结合自然语言任务描述、标量分数以及当可用时的结构化解释反馈。在包含23个任务的实用分子优化基准上,SEISMO在优化曲线下的面积比先前方法高2-3倍,通常在50次预言机调用内达到接近最大任务分数。我们额外的药物化学任务表明,提供解释反馈进一步提高了效率,证明了利用领域知识和结构化信息对于样本高效的分子优化至关重要。

英文摘要:

Optimizing molecules to achieve desired properties is a central bottleneck across the chemical sciences, particularly in the pharmaceutical industry, where it underlies the discovery of new drugs. Since molecular property evaluation often relies on costly and rate-limited oracles, such as experimental assays, molecular optimization must be highly sample-efficient. To address this, we introduce SEISMO, an LLM agent for inference-time molecular optimisation that turns information routinely available alongside the oracle score, but discarded by existing methods, into an explicit guidance signal. Rather than treating the oracle as a scalar black box, SEISMO conditions each proposal on a natural-language task description, the full optimization trajectory, and machine-readable feedback derived from post-hoc explainability methods and sub-score decompositions. Across a wide range of drug-discovery-relevant tasks, this consistently improves sample efficiency over existing optimisers as well as zero-shot LLM generation, with gains growing as explanatory feedback is enriched. In practice, medicinal chemists can inspect the agent's reasoning and intervene to steer generation in natural language, keeping them central to molecular optimisation projects.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑