Huan Ma | Tianjin University

About Me

I am currently a PhD student at Tianjin University, and my supervisor is Professor Changqing Zhang. I’m now focusing on uncertainty-aware machine learning related to LLMs.

News

[Aug. 2025] New research is released about Uncertainty Estimation in LLMs (Semantic Energy).
[Mar. 2025] New research is released about Uncertainty Estimation in LLMs (Token-level).
[Dec. 2024] Two papers about Visual-Language Foundation Models are accepted to AAAI 2025.
[Sept. 2023] Our paper about LLM is accepted to NeurIPS 2023.
[June 2023] Master’s thesis was awarded as an outstanding thesis at Tianjin University.
[May 2023] Our paper about multimodal learning is accepted to ICML 2023.
[Sept. 2021] Our paper about trustworthy multimodal regression is accepted to NeurIPS 2021.

Publications

arXiv

Semantic Energy: Detecting LLM Hallucination Beyond Entropy

Huan Ma, Jiadong Pan, Jing Liu, Yan Chen, Joey Tianyi Zhou, Guangyu Wang, Qinghua Hu, Hua Wu, Changqing Zhang, Haifeng Wang

arXiv preprint arXiv:2508.14496, 2025.

PDF Code BibTex

arXiv

Estimating LLM Uncertainty with Evidence

Huan Ma, Jingdong Chen, Joey Tianyi Zhou, Guangyu Wang, Changqing Zhang

arXiv preprint arXiv:2502.00290, 2025.

PDF Code BibTex

ICML(Oral)

Calibrating Multimodal Learning

Huan Ma, Qingyang Zhang, Changqing Zhang, Bingzhe Wu, Huazhu Fu, Joey Tianyi Zhou, Qinghua Hu

International Conference on Machine Learning (ICML), 2023.

PDF Code BibTex Oral Presentation, Primary Co-First Author

NeurIPS

Fairness-guided Few-shot Prompting for Large Language Models

Huan Ma, Changqing Zhang*, Yatao Bian, Lemao Liu, Zhirui Zhang, Peilin Zhao, Shu Zhang, Huazhu Fu, Qinghua Hu, Bingzhe Wu

Advances in Neural Information Processing Systems (NeurIPS), 2023.

PDF Code BibTex First Author

NeurIPS

Trustworthy Multimodal Regression with Mixture of Normal-inverse Gamma Distributions

Huan Ma, Zongbo Han, Changqing Zhang, Huazhu Fu, Joey Tianyi Zhou and Qinghua Hu

Advances in Neural Information Processing Systems (NeurIPS), 2021.

PDF Code BibTex Primary Co-First Author

AAAI

Spurious Feature Eraser: Stabilizing Test-Time Adaptation for Vision-Language Foundation Model

Huan Ma, Yan Zhu, Changqing Zhang, Peilin Zhao, Baoyuan Wu, Long-Kai Huang, Qinghua Hu, Bingzhe Wu

The Association for the Advancement of Artiffcial Intelligence (AAAI), 2025.

PDF Code BibTex Primary Co-First Author

AAAI

Exploring Task-Level Optimal Prompts for Visual In-Context Learning

Yan Zhu, Huan Ma, Changqing Zhang

The Association for the Advancement of Artiffcial Intelligence (AAAI), 2025.

PDF Code BibTex Secondary Co-First Author

Internship

ERNIE, Baidu

April 2025 -- Now

Research on LLM Uncertainty Estimation and Applications.

AI Lab, Tencent

April 2022 -- Nov. 2024

5 papers have been completed, and 3 of them have been accepted.
Built LLM content moderation system (harmful content detection) based on multi-agents.