Program analysis is an important class of techniques which are very helpful in developing dependable software systems. In particular, static program analysis tries to infer properties of the program without executing it. In this century, static program analysis tools are being used by more and more people. They helped developers and test engineers find intricate bugs successfully. Yet there are still some problems with such tools, such as scalability, efficiency and accuracy. In this talk, I will describe some techniques for static program analysis, in particular, our work on symbolic execution of C/C++ programs, and some of our recent attempts on using LLMs for static analysis tool development.
Professor
Institute of Software, Chinese Academy of Sciences, China
Jian Zhang is a research professor with the Institute of Software, Chinese Academy of Sciences (ISCAS), and a professor with the University of Chinese Academy of Sciences (UCAS). His research interests include software testing, program analysis, automated reasoning and constraint solving. He has served on the program committees of well-known international conferences such as ICSE, ISSTA, ISSRE, VSTTE, ICFEM and QSIC/QRS. He also serves on the editorial boards of several journals including Frontiers of Computer Science, Journal of Computer Science and Technology, and Chinese Journal of Computers.
