arXiv cs.AI· Zhentao Xu·· 3 天前AI 评分22
PID 门控的持续提示词自适应：噪声反馈下的鲁棒调度
PID-Gated Continual Prompt Adaptation Under Noisy Feedback
AI 导读
针对 LLM 智能体在非平稳环境中因噪声反馈导致指令更新被污染的问题，研究者受 PID 控制启发提出一族轻量门控机制，依据当前误差、历史证据累积与编辑后回滚来决定何时触发持久指令更新。在多种漂移任务、不同 LLM 参数规模与反馈噪声水平下，PID 门控在任务性能与编辑器调用预算之间取得比 TextGrad、GEPA 及简单基线更优的 Pareto 权衡。
正文 · 原文
Abstract:Deploying LLM agents in non-stationary environments requires continually revising their instructions as the task drifts, yet rewriting after every reported mistake can turn noisy feedback into persistent errors. This raises an often-overlooked scheduling question in prompt adaptation: when should feedback trigger a lasting instruction update, and at what editor-call cost? Taking inspiration from proportional-integral-derivative (PID) control, we introduce a family of lightweight gates that react to the current error, accumulate evidence over time, and roll back deterioration after an edit. We benchmark these gates against both simple baselines and advanced prompt optimizers (TextGrad and GEPA), over a diverse suite of drifting tasks, LLM parameter sizes, feedback-noise levels, and random seeds. PID-gated continual prompt adaptation yields a more favorable Pareto trade-off between task performance (high accuracy, low error) and editor-call budget than the baselines considered, pointing to control-theoretic scheduling as a simple, general tool for robust prompt adaptation under unreliable feedback.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.16129 [cs.LG] |
| (or arXiv:2610.16129v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.16129 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Zhentao Xu [view email]
[v1]
Mon, 12 Oct 2026 03:46:35 UTC (254 KB)
来源：arXiv cs.AI · arxiv.org