发表机构
Wuhan University of Technology（武汉理工大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
该研究针对LLM智能体多目标材料发现中局部优化困难的问题，提出TRACE框架，通过记录编辑过渡聚合证据估计编辑效果，在对比实验中提升了宏平均命中率。
AI 中文摘要
基于大语言模型（LLM）智能体的多目标材料发现，不仅受限于可提出候选材料的数量，还受限于每次成本高昂的性能评估对后续搜索步骤的有效指导程度。现有智能体主要存储已评估候选材料及其得分，因此它们知道哪些材料成功了，但不知道哪些可执行的编辑导致了有用的性能变化。当目标相互竞争时，这使得局部优化变得困难，因为一个能提升某一性能的编辑可能会损害另一性能。我们提出TRACE，一种过渡感知残差控制框架，它将已评估的编辑作为反馈的基本单位。TRACE将每个局部优化记录为带有观测性能差值的父-编辑-子过渡，聚合过渡证据以估计可复用的编辑效果，并根据未来编辑减少当前候选材料剩余约束违反的预测能力对其进行排序，同时避免损害已满足的目标。在相同主干的受控对比实验中，TRACE超越了最先进的LLM智能体基线LLEMA，将宏平均命中率从18.13%提升至25.96%。
英文摘要
Multi-objective materials discovery with LLM agents is often limited not only by how many candidates can be proposed, but by how effectively each costly property evaluation informs the next search step. Existing agents mainly store evaluated candidates and their scores, so they know which materials succeeded but not which executable edits caused useful property changes. This makes local refinement difficult when objectives compete and an edit that improves one property may damage another. We propose TRACE, a transition-aware residual control framework that treats evaluated edits as the basic unit of feedback. TRACE records each local refinement as a parent-edit-child transition with observed property deltas, aggregates transition evidence to estimate reusable edit effects, and ranks future edits by their predicted ability to reduce the current candidate's remaining constraint violations while avoiding damage to already satisfied objectives. In a controlled same-backbone comparison, TRACE improves over LLEMA, the state-of-the-art LLM-agent baseline, raising macro-average hit rate from 18.13\% to 25.96\%.
