Conformal Path Reasoning: Trustworthy Knowledge Graph Question Answering via Path-Level Calibration
Shuhang Lin, Chuhao Zhou, Xiao Lin, Zihan Dong, Kuan Lu, Zhencan Peng, Jie Yin, Dimitris N. Metaxas · 2026
Read Conformal Path Reasoning: Trustworthy Knowledge Graph Question Answering via Path-Level Calibration by Shuhang Lin, Chuhao Zhou, Xiao Lin, Zihan Dong, Kuan Lu, Zhencan Peng, Jie Yin, Dimitris N. Metaxas (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-08077.txt.
Description
Shuhang Lin and coauthors make knowledge graph question answering more trustworthy. Graph-based answers are grounded and easy to inspect, but systems rarely guarantee how often the true answer sits inside the returned set. Conformal prediction can give such guarantees, yet earlier attempts broke the coverage promise or returned sets that were far too large.
Their framework, Conformal Path Reasoning, calibrates scores at the level of whole reasoning paths and at the query level, which keeps the statistical assumptions valid. A small value network, trained with guided exploration, learns sharper path scores. Experiments show better empirical coverage and smaller, more useful answer sets.
Catalog record
| CREATOR | Shuhang Lin, Chuhao Zhou, Xiao Lin, Zihan Dong, Kuan Lu, Zhencan Peng, Jie Yin, Dimitris N. Metaxas |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49 KB |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.08077