Qiang Sheng(盛强)
Associate Professor / Researcher
Institute of Computing Technology, Chinese Academy of Sciences
Highlights
About
I am an Associate Professor/Researcher at the Media Synthesis and Forensics Lab, Institute of Computing Technology, Chinese Academy of Sciences. I am also a supervisor at School of Computer Science and Technology, University of Chinese Academy of Sciences (UCAS), where I got my Ph.D. degree under the supervision of Professor Juan Cao. My research interests include:
- Application: Fake News/Misinformation Detection, Fact-Checking, Machine-Generated Content Detection
- Direction: Natural Language Understanding and Social Media Mining
- Vision: Make the World More Credible
News
View All →One co-authored paper got accepted by ICML 2026.
Two co-authored papers got accepted by ACL 2026.
One co-authored paper got accepted by IEEE TVCG.
One co-authored paper got accepted by EACL 2026.
Will serve as a Senior PC Member at ICMR 2026.
Selected Publications
View All →ICML'26FactGuard: 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
Proceedings of the 43rd International Conference on Machine Learning (Acceptance Rate: 6352/23918=26.6%)
We propose an agentic misinformation video detection framework that can reason the veracity iteratively with self-refinement.
ACL'26Logical Consistency as a Bridge: Improving LLM Hallucination Detection via Label Constraint Modeling between Responses and Self-Judgments
Hao Mi, Qiang Sheng, Shaofei Wang, Beizhe Hu, Yifan Sun, Zhengjia Wang, Hengqi Zeng, Yang Li, Danding Wang, Juan Cao
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Acceptance Rate: 19%)
We model the logical label constraint between LLM responses and self-judgments as a bridge to enhance hallucination detection.
ACL'26Beyond the Final Actor: Modeling the Dual Roles of Creator and Editor for Fine-Grained LLM-Generated Text Detection
Yang Li, Qiang Sheng, Zhengjia Wang, Yehan Yang, Danding Wang, Juan Cao
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Acceptance Rate: 19%)
We disentangle the dual roles of creator and editor in LLM text generation, enabling fine-grained detection of AI-generated content at different revision modes.
IEEE TVCGInteractive Visual Assessment for Text-to-Image Generation Models
Xiaoyue Mi, Fan Tang, Juan Cao, Qiang Sheng, Ziyao Huang, Peng Li, Yang Liu, Tong-Yee Lee
IEEE Transactions on Visualization and Computer Graphics
We build an interactive visual assessment tool for exposing T2I models' vulnerability.
AAAI'26Reasoning About the Unsaid: Misinformation Detection with Omission-Aware Graph Inference
Zhengjia Wang, Danding Wang, Qiang Sheng, Jiaying Wu, Juan Cao
Proceedings of the 40th AAAI Conference on Artificial Intelligence (Acceptance Rate: 4167/23680=17.60%)
We consider the omitted information to better reason the creator's intent for misinformation detection.
NeurIPS'25From Judgment to Interference: Early Stopping LLM Harmful Outputs via Streaming Content Monitoring
Yang Li, Qiang Sheng, Yehan Yang, Xueyao Zhang, Juan Cao
Proceedings of the 39th Annual Conference on Neural Information Processing Systems (Acceptance Rate: 5290/21575=24.52%)
We build a content moderator that can early stop LLMs' harmful outputs with low latency.
CIKM'25Bridging Thoughts and Words: Graph-Based Intent-Semantic Joint Learning for Fake News Detection
Zhengjia Wang, Qiang Sheng, Danding Wang, Beizhe Hu, Juan Cao
Proceedings of the 34th ACM International Conference on Information and Knowledge Management (Acceptance Rate: 443/1627=27.2%)
We propose to inject intent information with graph-based joint learning into fake news detection.
CIKM'25Enhancing Fake News Video Detection via LLM-Driven Creative Process Simulation
Yuyan Bu, Qiang Sheng, Juan Cao, Shaofei Wang, Peng Qi, Yuhui Shi, Beizhe Hu
Proceedings of the 34th ACM International Conference on Information and Knowledge Management (Acceptance Rate: 185/604=30.6%)
We propose AgentAug, which simulates the creative process of fake news videos using LLMs to mitigate the data scarcity issue.
IP&MExploring News Intent and Its Application: A Theory-Driven Approach
Zhengjia Wang, Danding Wang, Qiang Sheng, Juan Cao, Siyuan Ma, Haonan Cheng
Information Processing & Management
We, for the first time, conceptualize and computerize news intent modeling and showcase its application on fake news detection, popularity prediction, and propaganda detection.
SIGIR'25LLM-Generated Fake News Induces Truth Decay in News Ecosystem: A Case Study on Neural News Recommendation
Beizhe Hu, Qiang Sheng, Juan Cao, Yang Li, Danding Wang
Proceedings of The 48th International ACM SIGIR Conference on Research and Development in Information Retrieval (Acceptance Rate: 239/1071=22.3%)
We reveal the truth-decay phenomenon where real news gradually loses its top-ranked advantage against fake news when LLM-generated news penetrates.