Research
My research work is at an early stage and focuses on the intersection of geometric data, machine learning, and AI safety. The projects below are undergraduate innovation projects conducted under the supervision of Prof. Xiuping Liu.
Undergraduate Research
Intelligent Representation and Safe Learning for Geometric Data
Led a provincial-level undergraduate innovation project investigating intelligent representations of geometric data and their connections to safe learning. The project covered research planning, technical exploration, and project coordination, and passed expert review and completion assessment.
Advisor: Prof. Xiuping Liu
Theoretical Research and Practice for AI Safety
A newly proposed project that aims to connect theoretical questions in AI safety with practical investigation. The planned work will examine mathematical and computational perspectives on model safety, robustness, and dependable learning.
Advisor: Prof. Xiuping Liu
Current Directions
AI Safety and Robustness
Studying adversarial machine learning and methods for building more reliable models.
3D Generative Modeling
Exploring diffusion-based generation, backdoor robustness, and representation learning for 3D point clouds.
Mathematical Foundations
Strengthening foundations in probability, statistics, numerical analysis, and optimization.