证据图结构何时预测答案正确性？结构化检索的无标签弃答
原标题：When Does Evidence-Graph Structure Predict Answer Correctness? Label-Free Abstention for Structured Retrieval
AI 摘要
该论文提出一种无需训练的结构化检索弃答方法，利用证据图中已选证据与未选候选之间的加权边界割（boundary cut）作为置信信号，无需正确性标签或置信模型训练。在 MemoryAgentBench 多跳任务上，该方法将精确率-覆盖率曲线下面积从最佳分数基线的水平提升至更高值（95% CI），零割子集包含的查询全部正确。研究进一步表明该信号的有效性取决于图边语义是否与答案支持兼容性对齐，类型化结构本身并不保证有用置信度。
正文节选
When Does Evidence-Graph Structure Predict Answer Correctness? Label-Free Abstention for Structured Retrieval Abstract Retrieval systems return scored answers, but a high score does not show that an answer should be trusted. The proposed training-free abstention method applies the boundary cut to the selected evidence in an evidence graph. The signal is the weighted cut between selected and unselected candidates and requires no correctness labels or confidence-model training. On MemoryAgentBench