I am a final-year Ph.D. Candidate at UCLA Computer Science Department, advised by Prof. Cho-Jui Hsieh. My primary research focus is on trustworthy machine learning, where I study formal methods for verifying and training ML models with verifiable safety guarantees in mission-critical applications, as well as empirical methods for the robustness and safety evaluation and defense for large-scale ML foundation models. I received my B.Eng. degree from the CST department at Tsinghua University.
I am currently on the job market.
Formal Verification for ML: General and scalable approaches for formally verifying NNs with diverse model architectures and verification specifications, to enable automatic verification in a broad range of ML applications.
Training Verifiably Robust and Safe ML Models: Verification-aware NN training methods to efficiently train NNs with stronger verifiable safety guarantees.
Verifiably Safe NN-based Control Systems: Applications of both formal verification and verification-aware NN training to the safety of NN-based control systems.
Empirical Robustness Evaluation and Defense for ML Foundation Models: Empirical methods for evaluating and enhancing the robustness and safety of large-scale ML foundation models such as LLMs.
TA at UCLA: