Qi Wu

I am an undergraduate student at the University of Science and Technology of China (USTC). My research interests sit at the intersection of multimodal learning, reliable machine learning, and medical artificial intelligence.

I am particularly interested in trustworthy multimodal intelligence: models that can connect visual and linguistic evidence, express uncertainty appropriately, and support careful evaluation in real-world applications.

Research interests

• Multimodal Intelligence — reasoning, perception, and representation in vision-language models and multimodal large language models.
• Trustworthy AI — uncertainty estimation, calibration, evaluation, robustness, and the reliability of modern AI systems.
• AI for Healthcare — responsible development and evaluation of multimodal AI systems for high-stakes medical applications.

Research experience

Trustworthy Medical Vision-Language Models — USTC
Advisor: Prof. Qingsong Yao · January 2026–Present
Projects: Counterfactual Answer Sensitivity for Reliable Medical VLM Evaluation; ClinEviD: Decision-Token Evidence Discrepancy for Medical VQA Uncertainty Quantification

The Exploration of Preference Optimization in LLMs for Recommendation Systems — USTC
Advisor: Prof. Xiangnan He · January–June 2025

Education

University of Science and Technology of China (USTC)
Undergraduate Student, 2023–Present

University of California, Berkeley (UC Berkeley)
Exchange Undergraduate Student, January–May 2026

Beyond research

Outside research, I maintain a personal blog about books, films, literature, and personal reflections. I also love music: I began studying classical piano at six and play keyboard in a campus rock band.