RSI-PDEbench 基准发布：评估 LLM 智能体求解偏微分方程的能力
RSI-PDEbench: A Benchmark for Self-Improving Agents in PDE Solving
把“从数据反推 PDE 方程”这件原本要专家手工做的事交给 LLM 智能体，还做了个 1,001 案例的基准来量化。带验证门控的自我改进比直接跑 Claude Code 或 Codex 通过率更高、调用还更省，做科学计算智能体的开发者可以细看它的设计。
arXiv 论文 2611.14949 提出 RSI-PDEbench 基准，包含 1,001 个案例、覆盖 42 个方程模板，考察 LLM 智能体从采样数据中识别并求解偏微分方程的完整流程。基准设置三种评测：Full 用无噪声数据测基础性能；Noisy 注入 RMS 等于干净信号 50% 的高斯噪声测鲁棒性；Hard-OOD 要求泛化到未见过的方程模板和偏移的系数范围。
论文同时推出 RSI-PDE，一个递归自我改进的智能体框架：先按残差损失筛选候选方程，再用相对 L² 反馈迭代修正系数与数值求解器，最后通过验证门控决定策略和记忆更新是否生效。相比单次运行基线，RSI-PDE 在四个前沿模型上的平均通过率提升 106.5%–181.9%；相比 Claude Code 或 Codex 原生框架提升 3.0%–19.6%，同时把 12 组模型-设置组合的模型调用总量减少 46.9%。
RSI-PDEbench: A Benchmark for Self-Improving Agents in PDE Solving
Identifying and solving partial differential equations (PDEs) from sampled solution data often requires substantial expert effort in equation discovery, coefficient estimation, and numerical solver configuration, especially under noise and distribution shifts. Can large language model (LLM) agents automate this workflow accurately and efficiently? To investigate this question, we introduce \bench{}, a benchmark of 1,001 cases spanning 42 equation templates, including both classical and fractional PDEs. It provides three evaluation settings: Full evaluates performance on noise-free solution data; Noisy tests robustness to Gaussian noise with an RMS equal to 50\% of the clean-signal RMS; and Hard-OOD tests challenging out-of-distribution generalization to unseen equation templates and shifted coefficient ranges. We further introduce RSI-PDE, a recursively self-improving agent harness that shortlists candidate equations by residual loss, uses relative $L^2$ feedback to refine coefficients and numerical solvers, and applies a validation gate to policy and memory updates. We compare RSI-PDE with single-run baselines and native agent harnesses (Claude Code or Codex) across four frontier models. Across the three settings, RSI-PDE improves the pass rate averaged across the four models by 106.5--181.9\% relative to single-run baselines and by 3.0--19.6\% relative to native harnesses, while reducing total recorded model calls by 46.9\% across the 12 model--setting pairs. These results highlight the potential of validated recursive self-improvement in LLM agents to advance scientific computing.
本条情报归属 AITOP「论文」核心追踪栏目。在跨区域产业观察坐标中，该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线，反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括：RSI-PDEbench、LLM智能体、科学计算、偏微分方程，读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明：本文原始线索采编自「arXiv cs.AI」（@Binghang Lu, Lang Cao, Changhong Mou, Yige Sun, Cindy Xiangrui Kong, Zheyuan Deng, Tian Yuan, Bing Hu, Runyu Zhang, Na Lu, Guang Lin），经 AITOP 多维评分管线结构化提炼。