Weaver：通过增量推理编织偏好证据的个性化RAG框架
原标题：Weaver: Weaving Preference Evidence through Incremental Reasoning for Personalized Retrieval-Augmented Generation
AI 摘要
该论文提出 Weaver，一个面向个性化检索增强生成（RAG）的框架，通过双层检索与推理图（上层意图依赖图、下层记忆条件检索轨迹图）增量地整合分散在历史对话中的用户偏好证据。其核心模块包括依赖信息缺口追踪、记忆条件增量检索和范围感知组合推理，用于区分互补、替代与冲突的偏好证据。在 PersonaMem 和 PrefEval 基准上，Weaver 分别达到 86.42% 和 91.10% 的准确率，超过最强基线 16.01 和 7.56 个百分点。
正文节选
Weaver: Weaving Preference Evidence through Incremental Reasoning for Personalized Retrieval-Augmented Generation Abstract. Personalized retrieval-augmented generation requires extracting and applying user preferences from extended interaction histories. This task is challenging because preference evidence is scattered across past conversations and remains bound to specific temporal and situational contexts. Consequently, isolated fragments rarely provide complete support for a response, while p