Papers
9
Total Citations
473
H-Index
6
About
Weiming Xiang is a prominent researcher specializing in the formal verification and safety analysis of neural networks and cyber-physical systems. His work sits at the critical intersection of machine learning, control theory, and formal methods, addressing one of the most pressing challenges in modern AI: ensuring the reliability of neural network-based systems in safety-critical applications. Xiang is perhaps best known for pioneering reachability analysis techniques for neural networks. His foundational 2018 paper on output reachable set estimation for multilayer perceptrons, which introduced the influential concept of "maximum sensitivity," has accumulated over 270 citations and established a cornerstone methodology in neural network verification. Building on this, his 2020 simulation-guided reachability approach for neural network control systems further demonstrated how AI vulnerabilities in cyber-physical systems can be rigorously characterized and mitigated. Beyond verification, Xiang has contributed to parallelizable reachability algorithms, robust training frameworks grounded in formal methods, and distributed neural hybrid system models for complex dynamical systems. His early work on hybrid systems verification tools also reflects a broad expertise in formal analysis across multiple computational paradigms. With hundreds of citations across his portfolio, Xiang's research has meaningfully advanced the theoretical and practical foundations for deploying trustworthy AI in real-world safety-critical environments.
Research Focus
Key Achievements
Top Papers
- 1Output Reachable Set Estimation and Verification for Multilayer Neural Networks270 citations · 2018
- 2
- 3
- 4
- 5
- 6
- 7
- 8
- 9