FactGuard: Agentic Video Misinformation Detection via Reinforcement Learning

Zehao Li, Hongwei Yu, Hao Jiang, Qiang Sheng, Yilong Xu, Baolong Bi, Yang Li, Zhenlong Yuan, Yujun Cai, Zhaoqi Wang · 2026

Read FactGuard: Agentic Video Misinformation Detection via Reinforcement Learning by Zehao Li, Hongwei Yu, Hao Jiang, Qiang Sheng, Yilong Xu, Baolong Bi, Yang Li, Zhenlong Yuan, Yujun Cai, Zhaoqi Wang (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-2602-22963.txt.

Description

Zehao Li and nine co-authors build FactGuard, an agentic checker for video misinformation. Large multimodal models often reason at fixed depth and trust their own guesses, even when key evidence is thin, split, or needs outside checks. FactGuard instead judges task ambiguity and calls outside tools for missing proof, refining its reasoning step by step.

A two-stage training recipe pairs domain-specific agentic fine-tuning with decision-aware reinforcement learning, tuning tool use and risk-sensitive choices. Tests on FakeSV, FakeTT, and FakeVV show state-of-the-art results with strong robustness and generalization, a step toward video fact-checking that knows when to look outward.

Catalog record

CREATORZehao Li, Hongwei Yu, Hao Jiang, Qiang Sheng, Yilong Xu, Baolong Bi, Yang Li, Zhenlong Yuan, Yujun Cai, Zhaoqi Wang
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE47 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2602-22963

Browse the full library · About the Tower of Babel