arXiv cs.AI· Haosen Zhang, Yang Yang·· 10 小时前AI 评分33
EASER：弥合证据到执行鸿沟的多目标肽设计反思型智能体
Bridging the Evidence-to-Execution Gap:A Reflective Agent for Multi-Objective Peptide Design
AI 导读
EASER 是一个反思型智能体，通过可学习的属性接口把文献推理与序列生成连接起来：LLM 控制器组合离线训练的低秩方向，操控冻结的扩散生成器，并以 Probe-and-Steer 检验干预假设、通过反思更新记忆。在计算抗菌肽设计任务上，EASER 平衡活性、非溶血性与非毒性，显式假设构建带来的多目标增益优于直接生成动作，且在筛选候选上取得最高平均 hypervolume 与最低平均 IGD+。
正文 · 原文
Abstract:Large language models can reason over scientific literature, but textual plans do not reliably produce biological sequences with the intended properties. Sequence generators capture sequence patterns, yet they do not interpret evidence, choose interventions, or reflect on outcomes. This separation creates an evidence-to-execution gap between scientific reasoning and sequence-level control. We introduce EASER, a reflective agent that connects the two through a learned property interface. The interface consists of property-specific low-rank directions trained offline and fixed during search; an LLM controller (the controller) composes them to steer a frozen diffusion generator. From retrieved evidence, sequence context, and prior outcomes, it proposes intervention hypotheses that specify anchors, editable positions, and coefficients. Probe-and-Steer tests these hypotheses and allocates further samples from their joint predicted responses; reflection updates the agent's memory. On computational antimicrobial peptide design, EASER balances activity, non-hemolysis, and non-toxicity. Explicit hypothesis formulation yields larger multi-objective gains than direct action generation, and ablations support evidence retrieval, episodic history, reflection, and Probe-and-Steer. EASER achieves the highest mean hypervolume and lowest mean IGD+ among the evaluated search policies and generators on screened candidates. These results show that a learned property interface can make literature-grounded reasoning executable at the sequence level.
| Comments: | 24 pages, 8 figures |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.06190 [cs.AI] |
| (or arXiv:2610.06190v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.06190 arXiv-issued DOI via DataCite |
Submission history
From: Haosen Zhang [view email]
[v1]
Mon, 5 Oct 2026 12:07:21 UTC (3,563 KB)
[v2]
Tue, 10 Nov 2026 08:57:01 UTC (2,874 KB)
来源：arXiv cs.AI · arxiv.org