Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2604.19005 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866917424863903744 |
|---|---|
| author | Tang, Yixuan Zhang, Yirui Feng, Hang Tung, Anthony K. H. |
| author_facet | Tang, Yixuan Zhang, Yirui Feng, Hang Tung, Anthony K. H. |
| contents | Half-truths, claims that are factually correct yet misleading due to omitted context, remain a blind spot for fact verification systems focused on explicit falsehoods. Addressing such omission-based manipulation requires reasoning not only about what is said, but also about what is left unsaid. We propose RADAR, a role-anchored multi-agent debate framework for omission-aware fact verification under realistic, noisy retrieval. RADAR assigns complementary roles to a Politician and a Scientist, who reason adversarially over shared retrieved evidence, moderated by a neutral Judge. A dual-threshold early termination controller adaptively decides when sufficient reasoning has been reached to issue a verdict. Experiments show that RADAR consistently outperforms strong single- and multi-agent baselines across datasets and backbones, improving omission detection accuracy while reducing reasoning cost. These results demonstrate that role-anchored, retrieval-grounded debate with adaptive control is an effective and scalable framework for uncovering missing context in fact verification. The code is available at https://github.com/tangyixuan/RADAR. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_19005 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Debating the Unspoken: Role-Anchored Multi-Agent Reasoning for Half-Truth Detection Tang, Yixuan Zhang, Yirui Feng, Hang Tung, Anthony K. H. Computation and Language Half-truths, claims that are factually correct yet misleading due to omitted context, remain a blind spot for fact verification systems focused on explicit falsehoods. Addressing such omission-based manipulation requires reasoning not only about what is said, but also about what is left unsaid. We propose RADAR, a role-anchored multi-agent debate framework for omission-aware fact verification under realistic, noisy retrieval. RADAR assigns complementary roles to a Politician and a Scientist, who reason adversarially over shared retrieved evidence, moderated by a neutral Judge. A dual-threshold early termination controller adaptively decides when sufficient reasoning has been reached to issue a verdict. Experiments show that RADAR consistently outperforms strong single- and multi-agent baselines across datasets and backbones, improving omission detection accuracy while reducing reasoning cost. These results demonstrate that role-anchored, retrieval-grounded debate with adaptive control is an effective and scalable framework for uncovering missing context in fact verification. The code is available at https://github.com/tangyixuan/RADAR. |
| title | Debating the Unspoken: Role-Anchored Multi-Agent Reasoning for Half-Truth Detection |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2604.19005 |