[论文] Conformal Path Reasoning: Trustworthy Knowledge Graph Question Answeri...

## 论文概要 **研究领域**: NLP **作者**: Shuhang Lin, Chuhao Zhou, Xiao Lin **发布时间**: 2025-05-07 **arXiv**: [2505.05130](https://arxiv.org/abs/2505.05130) ## 中文摘要 知识图谱问答（KGQA）在基于实地的可解释推理方面显示出前景，但现有方法往往无法为检索到的答案提供可靠的覆盖保证。虽然共形预测（CP）为生成具有统计保证的预测集提供了一个有原则的框架，但先前方法在校准有效性和得分可区分性方面存在关键限制，导致覆盖保证被破坏和预测集过大。为解决这些陷阱，我们提出共形路径推理（CPR），一个可信的KGQA框架，包含两项关键创新。首先，我们在路径级得分上执行查询级共形校准，保留可交换性同时生成路径预测集。其次，我们引入残差共形值网络（RCVNet），一个轻量级模块，通过PUCT引导探索训练以学习有区分性的路径级非共形得分。基准上的实验表明，与共形基线相比，CPR显著提高了经验覆盖率34%，同时将平均预测集大小减少40%。这些结果验证了CPR在满足覆盖保证的同时获得更紧凑答案集的有效性。 ## 原文摘要 Knowledge Graph Question Answering (KGQA) has shown promise for grounded and interpretable reasoning, yet existing approaches often fail to provide reliable coverage guarantees over retrieved answers. While Conformal Prediction (CP) offers a principled framework for producing prediction sets with statistical guarantees, prior methods suffer from critical limitations in both calibration validity and score discriminability, resulting in violated coverage guarantees and excessively large prediction sets. To address these pitfalls, we propose Conformal Path Reasoning (CPR), a trustworthy KGQA framework with two key innovations. First, we perform query-level conformal calibration over path-level scores, preserving the exchangeability while generating path prediction sets. Second, we introduce t... --- *自动采集于 2026-05-12* #论文 #arXiv #NLP #小凯