🧑‍🎓About Me

My current research interests center on Trustworthy AI, specifically focusing on Uncertainty Quantification and Adversarial Robustness. I am open to any form of academic cooperation, please feel free to email me at [thuang[at]bjtu[dot]edu[dot]cn].

I received my B.S. degree from the School of Mathematics and Statistics, Beijing Jiaotong University (北京交通大学数学与统计学院), and was directly admitted to the Ph.D. program at the School of Computer Science and Technology, Beijing Jiaotong University (北京交通大学计算机科学与技术学院), advised by Liping Jing (景丽萍). I also collaborate with Rui Wang (王睿) and Huafeng Liu (刘华锋).

I have published several papers at top-tier international AI conferences such as ICLR, CVPR, ICCV.

For more details on my research and academic experience, please refer to my Curriculum Vitae.

🔥 News

• 2026.02: 🎉 One paper on Uncertainty Quantification was accepted by ICLR 2026.
• 2025.12: 🚀 Our dataset PRE-HAL reached 13k+ downloads on Hugging Face!
• 2025.07: 🎉 One paper on SAM Safety was accepted by ICCV 2025.
• 2025.06: 🎉 One paper on Hallucination Detection was accepted by IJAR 2025.
• 2024.02: 🎉 One paper on Adversarial Attack was accepted by CVPR 2024.
• 2023.09: 🎓 Started my Ph.D. at Beijing Jiaotong University.

📚 Publications

📄 Paper List

📊 Uncertainty Quantification

• ICLR 2026 Detecting Misbehaviors of Large Vision-Language Models by Evidential Uncertainty Quantification, Tao Huang, Rui Wang, Xiaofei Liu, Yi Qin, Li Duan, Liping Jing.
• IJAR 2025 Visual Hallucination Detection in Large Vision-Language Models via Evidential Conflict, Tao Huang, Zhekun Liu, Rui Wang, Yang Zhang, Liping Jing. [🤗 Dataset:PRE-HAL]

🛡️ Adversarial Learning

• ICCV 2025 SAM Encoder Breach by Adversarial Simplicial Complex Triggers Downstream Model Failures, Yi Qin, Rui Wang, Tao Huang, Tong Xiao, Liping Jing.
• CVPR 2024 Strong Transferable Adversarial Attacks via Ensembled Asymptotically Normal Distribution Learning, Zhengwei Fang, Rui Wang, Tao Huang, Liping Jing.

⭐ Selected Publications

ICLR 2026

Detecting Misbehaviors of Large Vision-Language Models by Evidential Uncertainty Quantification

Tao Huang, Rui Wang, Xiaofei Liu, Yi Qin, Li Duan, Liping Jing; | 📦Code | 🤗Huggingface

Efficient Framework: A training-free method that quantifies uncertainty using pre-logits features in a single forward pass.

Novel Insight: Layer-wise dynamic analysis reveals hallucinations correspond to high internal conflict, while OOD failures correspond to high ignorance.

Superior Performance: Outperforms strong baselines on four state-of-the-art LVLMs, improving AUROC by 10.4% and AUPR by 5.3% on average.

📖 Educations

• 2023.09 – Present, Ph.D. Student, School of Computer Science and Technology, Beijing Jiaotong University, Beijing, China

• 2019.09 – 2023.06, B.Sc., School of Mathematics and Statistics, Beijing Jiaotong University, Beijing, China