用两阶段优化提升小样本关系抽取效果，先推理后微调。
Two-Stage Prompt Optimization for Few-Shot Relation Extraction: From Reasoning-Guided Search to Gradient-Guided Refinement
- 先用自然语言推理优化提示，再用梯度信号精修关键片段。
- 在FS-TACRED上达到当前最佳性能，比基线提升超过3个百分点。
- 适合小模型做少样本关系抽取，尤其适用于资源受限场景。
小规模语言模型在事件式少样本关系抽取中的自动提示优化仍不充分。本文提出一种两阶段框架，结合基于推理的提示优化与基于梯度的提示优化。第一阶段可使用任意基于推理的优化器，在自然语言层面实现提示的广泛改进；第二阶段采用我们提出的GradPO方法，利用损失和梯度信号识别高影响提示片段，并通过局部修改进行精细化调整。在FS-TACRED和FS-FewRel数据集上的实验表明，局部精修通常能提升第一阶段找到的提示表现，且GradPO是最具一致性的优化器。该框架在使用Qwen3-4B模型时，在FS-TACRED上达到当前最优性能，同时在FS-FewRel上保持竞争力。
原文摘要 · Abstract (English)
Automatic prompt optimization is still underexplored for episodic few-shot relation extraction with smaller language models. We propose a two-stage framework that combines reasoning-based prompt optimization with gradient-based prompt optimization. The first stage can use any reasoning-based optimizer to make broadprompt improvements in natural language. The second stage applies our GradPO, which uses loss and gradient signals to identify high-impact prompt spans and refine them with local edits. Experiments on FS-TACRED and FS-FewRel show that local refinement usually improves prompts found by the first stage, and GradPO is the most consistent refiner. Our framework achieves state-of-the-art performance on FS-TACRED with Qwen3-4B and remains competitive on FS-FewRel.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。