Conformal Path Reasoning: Trustworthy Knowledge Graph Question Answering via Path-Level Calibration

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lin, Shuhang, Zhou, Chuhao, Lin, Xiao, Dong, Zihan, Lu, Kuan, Peng, Zhencan, Yin, Jie, Metaxas, Dimitris N.
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909027279044608
author Lin, Shuhang
Zhou, Chuhao
Lin, Xiao
Dong, Zihan
Lu, Kuan
Peng, Zhencan
Yin, Jie
Metaxas, Dimitris N.
author_facet Lin, Shuhang
Zhou, Chuhao
Lin, Xiao
Dong, Zihan
Lu, Kuan
Peng, Zhencan
Yin, Jie
Metaxas, Dimitris N.
contents 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 the Residual Conformal Value Network (RCVNet), a lightweight module trained via PUCT-guided exploration to learn discriminative path-level nonconformity scores. Experiments on benchmarks show that CPR significantly improves the Empirical Coverage Rate by 34% while reducing average prediction set size by 40% compared to conformal baselines. These results validate the efficacy of CPR in satisfying coverage guarantees with substantially more compact answer sets.
format Preprint
id arxiv_https___arxiv_org_abs_2605_08077
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Conformal Path Reasoning: Trustworthy Knowledge Graph Question Answering via Path-Level Calibration
Lin, Shuhang
Zhou, Chuhao
Lin, Xiao
Dong, Zihan
Lu, Kuan
Peng, Zhencan
Yin, Jie
Metaxas, Dimitris N.
Computation and Language
68T05, 68T50, 62G15
I.2.4; I.2.7; I.2.6; G.3
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 the Residual Conformal Value Network (RCVNet), a lightweight module trained via PUCT-guided exploration to learn discriminative path-level nonconformity scores. Experiments on benchmarks show that CPR significantly improves the Empirical Coverage Rate by 34% while reducing average prediction set size by 40% compared to conformal baselines. These results validate the efficacy of CPR in satisfying coverage guarantees with substantially more compact answer sets.
title Conformal Path Reasoning: Trustworthy Knowledge Graph Question Answering via Path-Level Calibration
topic Computation and Language
68T05, 68T50, 62G15
I.2.4; I.2.7; I.2.6; G.3
url https://arxiv.org/abs/2605.08077