TaoSR1：电商相关性搜索的思考模型
TaoSR1: The Thinking Model for E-commerce Relevance Search
AI总结：
本文提出TaoSR1框架，通过CoT引导的监督微调、离线采样与DPO优化，解决电商搜索中相关性预测的推理误差与幻觉问题，实现高效部署。
摘要：
查询-商品相关性预测是电商搜索的核心任务。基于BERT的模型在语义匹配上表现优异，但缺乏复杂的推理能力。尽管大型语言模型（LLMs）被探索，大多数仍使用判别性微调或蒸馏到小模型进行部署。我们提出一个框架，直接部署LLMs用于此任务，解决关键挑战：推理链（CoT）误差累积、判别性幻觉和部署可行性。我们的框架TaoSR1包括三个阶段：（1）使用CoT的监督微调以培养推理能力；（2）离线采样与pass@N策略和直接偏好优化（DPO）以提高生成质量；（3）基于难度的动态采样与组相对策略优化（GRPO）以缓解判别性幻觉。此外，后CoT处理和基于累积概率的分区方法使在线部署高效。TaoSR1在离线数据集上显著优于基线，并在在线双人评估中取得显著优势，引入了将CoT推理应用于相关性分类的新范式。
英文摘要：
Query-product relevance prediction is a core task in e-commerce search. BERT-based models excel at semantic matching but lack complex reasoning capabilities. While Large Language Models (LLMs) are explored, most still use discriminative fine-tuning or distill to smaller models for deployment. We propose a framework to directly deploy LLMs for this task, addressing key challenges: Chain-of-Thought (CoT) error accumulation, discriminative hallucination, and deployment feasibility. Our framework, TaoSR1, involves three stages: (1) Supervised Fine-Tuning (SFT) with CoT to instill reasoning; (2) Offline sampling with a pass@N strategy and Direct Preference Optimization (DPO) to improve generation quality; and (3) Difficulty-based dynamic sampling with Group Relative Policy Optimization (GRPO) to mitigate discriminative hallucination. Additionally, post-CoT processing and a cumulative probability-based partitioning method enable efficient online deployment. TaoSR1 significantly outperforms baselines on offline datasets and achieves substantial gains in online side-by-side human evaluations, introducing a novel paradigm for applying CoT reasoning to relevance classification.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
