Xin Zhang

Hello, I am Xin Zhang (pronunciation ≈ "shin chang"), an assistant professor at Peking University. Before joining Peking, I was a postdoctoral associate at MIT CSAIL working with Prof. Armando Solar-Lezama. I received my Ph.D. from Georgia Tech under the supervision of Prof. Mayur Naik. I am broadly interested in topics related to programming languages and software engineering.

Over the past five years, my main research focus has been a new paradigm of program analysis that incorporates probabilistic reasoning into conventional abstract-interpretation-based program analysis. This paradigm, named Bayesian Program Analysis, enables program analyses to quantify results confidence and learn from external information. Around this paradigm, I have built applications in program analysis [OOPSLA'25a], fuzzing [POPL'26a], and fault localization [TSE'25,ICSE'22], developed algorithms for abstraction selection [OOPSLA'24a, OOPSLA'25b], question selection [OOPSLA'26], and efficient inference [ASE'25], and theories for abstract interpretation with confidence (conditionally accepted to PLDI'26). My other research interests include optimizing domain-specific languages for program synthesis [POPL'26b], artificial intelligence explainability [AAAI'25], autoformalization with LLMs [TOPLAS'26], and abstraction selection for traditional program analyses [SAS'21, OOPSLA'24b]. For details, please see Research.