ECPO : Evidence-Coupled Policy Optimization for Evidence-Certified Candidate Ranking
Miaobo Hu 1,2 Shuhao Hu 1,2 Bokun Wang 1,2 Yina Sa 1,2 Xiaobo Guo 1 Xin Wang 1 Daren Zha 1 Jun Xiao 3 1 Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China 2 School of Cyber Security, University of Chinese Academy of Sciences, Beijing, China 3 School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China [email protected] · 2026
Read ECPO : Evidence-Coupled Policy Optimization for Evidence-Certified Candidate Ranking by Miaobo Hu 1,2 Shuhao Hu 1,2 Bokun Wang 1,2 Yina Sa 1,2 Xiaobo Guo 1 Xin Wang 1 Daren Zha 1 Jun Xiao 3 1 Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China 2 School of Cyber Security, University of Chinese Academy of Sciences, Beijing, China 3 School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China [email protected] (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-2605-21993.txt.
Introduction
Miaobo Hu and co-authors present ECPO, evidence-coupled policy optimization for verifier-audited ranking. Ranking systems for decision support should expose evidence checkable apart from ranking scores. In a closed-roster setting, a policy emits a Top-K ranking jointly plus candidate-linked, span-grounded evidence bundles. ECPO combines listwise utility, certificate validity, and a reward that strips answer-bearing fields before deterministic recovery. MAVEN-ERE and RAMS tests show improved traceability, recovery that is rank-blind, and human support, with gains holding across held-out evaluators and backbones.
Catalog record
| CREATOR | Miaobo Hu 1,2 Shuhao Hu 1,2 Bokun Wang 1,2 Yina Sa 1,2 Xiaobo Guo 1 Xin Wang 1 Daren Zha 1 Jun Xiao 3 1 Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China 2 School of Cyber Security, University of Chinese Academy of Sciences, Beijing, China 3 School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China [email protected] |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 217515 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.21993