About
I am a Master's student in Intelligent Science and Technology at the University of Science and Technology of China (USTC), where I am a member of the Lab for Data Science (LDS). I am fortunate to be advised by Prof. Xiangnan He and Prof. Fuli Feng. I received my B.Eng. in Electronic Information Engineering from USTC in 2025.
My research focuses on making large language models more reliable and effective for reasoning and recommendation. I am particularly interested in reinforcement learning and preference optimization, uncertainty and calibration, and LLM-based recommender and multi-agent systems.
Outside research, I enjoy traveling, discovering local food, and listening to Mandopop.
News
Selected Publications
Full publication list →Uncertainty-aware Generative Recommendation
Chenxiao Fan, Chongming Gao, Yaxin Gong, Haoyan Liu, Fuli Feng, Xiangnan He
KDD 2026
Addresses uncertainty blindness in generative recommendation through uncertainty-weighted rewards, difficulty-aware optimization, and explicit confidence alignment.
Fine-grained List-wise Alignment for Generative Medication Recommendation
Chenxiao Fan, Chongming Gao, Wentao Shi, Yaxin Gong, Zihao Zhao, Fuli Feng
NeurIPS 2025
SpotlightFormulates medication recommendation as list-wise generation and uses step-wise GRPO to improve accuracy while controlling drug--drug interaction risk.
Process-Supervised LLM Recommenders via Flow-guided Tuning
Chongming Gao, Mengyao Gao, Chenxiao Fan, Shuai Yuan, Wentao Shi, Xiangnan He
SIGIR 2025
Uses GFlowNet-based process supervision to align token-level generation with item-level rewards, improving accuracy, diversity, and fairness.
Medical Reasoning With Large Language Models: A Systematic Review and Evaluation
Xiaohan Ren, Chenxiao Fan, Wenyin Ma, Hongliang He, Chongming Gao, Xiaoyan Zhao, Fuli Feng
iNew Medicine 2026
Systematizes LLM-based medical reasoning, evaluates representative models under a unified setting, and introduces MR-Bench using real-world clinical data.
Beyond Verbalized Confidence: Calibrating Reasoners with Differentiable Readouts
Chenxiao Fan, Chongming Gao, Gangyi Zhang, Leyang Shen, Yaxin Gong, Jiamin Wang, Jiakai Wang, Dong Wang, Yang Liu, Fuli Feng, Xiangnan He
arXiv 2026
Introduces CREDO, a differentiable confidence readout for RLVR that jointly improves reasoning accuracy, calibration, abstention, and selective prediction.
Trustworthy Recommendation in the Era of Large Language Models: Opportunities and Challenges
Bohao Wang, Yu Cui, Zhenxiang Xu, Jujia Zhao, Chenxiao Fan, Jizhi Zhang, Weiqin Yang, Shengjia Zhang, Sirui Chen, Yang Zhang, Xiaoyan Zhao, Wenjie Wang, Chongming Gao, Fuli Feng, Xiangnan He, Jiawei Chen
arXiv 2026
Reviews over 200 studies and maps 13 opportunities and 18 challenges for trustworthy LLM-empowered recommendation across six dimensions.
Fine-grained Alignment of Large Language Models for General Medication Recommendation without Overprescription
Zihao Zhao, Chenxiao Fan, Junlong Liu, Zheng Wang, Xiangnan He, Chongming Gao, Juan Li, Fuli Feng
arXiv 2025
Adapts LLMs to medication recommendation with medication-aware training, reducing overprescription and improving temporal and cross-dataset generalization.
Education
University of Science and Technology of China
Advisor: Prof. Xiangnan He
University of Science and Technology of China
Honors
First-Class Scholarship for Graduate Students
Outstanding Graduate of Anhui Province
Soong Ching Ling Scholarship
Scholarship for Talents in Information Technology
Excellent Student Scholarship
Excellent Student Scholarship