arXiv cs.AI· Hao Mao, Xu T. Liu, Peng Zhao, Shuai Lu, Yuntian Chen·· 12 小时前AI 评分34
LegoSR：面向智能体引导科学发现的统一自改进框架
LegoSR: A Unified Framework for Self-improving Agent-Guided Scientific Discovery
AI 导读
LegoSR 是一个统一代数符号回归与微分方程发现的框架，通过共享方程表示和评估流程整合了 7 个求解器，并支持部分已知方程约束下的搜索。其 LegoSR Agent 能根据求解器知识、中间结果和过往任务经验选择求解器并调整搜索，在含噪常微分与偏微分方程基准上恢复的方程数量接近 7 个专家调优求解器的并集，且比同模型、同工具、同预算的通用智能体高出最多 29 个百分点。代码已开源。
正文
Abstract:Equation discovery solvers often use different data interfaces, equation representations and evaluation criteria, making their results difficult to compare. We present LegoSR, a unified framework for algebraic symbolic regression and differential equation discovery. LegoSR integrates seven solvers through a common equation representation and evaluation process, and supports searches constrained by partially known equations. Researchers can run solvers directly or use LegoSR Agent, which selects solvers and adapts searches based on solver knowledge, intermediate results and lessons from earlier tasks. On noisy ordinary and partial differential equation benchmarks, LegoSR Agent recovers nearly as many equations as the union of seven expert-tuned solvers and outperforms a generic agent with the same model, tools and budget by up to 29 percentage points. Three examples demonstrate basic use, adaptive solver selection and discovery guided by partial physical knowledge.
| Comments: | 36 pages, 6 figures. Code is available at this https URL |
| Subjects: | Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.33465 [cs.AI] |
| (or arXiv:2610.33465v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2610.33465 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Hao Mao [view email]
[v1]
Sat, 24 Oct 2026 05:36:55 UTC (437 KB)
来源：arXiv cs.AI · arxiv.org