First Three Years of the International Verification of Neural Networks Competition (VNN-COMP)
It provides a meta-analysis for researchers and practitioners in AI safety and verification, but is incremental as it reviews existing competition data without introducing new methods.
This paper summarizes the first three years of the International Verification of Neural Networks Competition (VNN-COMP), which evaluates software tools for verifying safety and robustness properties of neural networks across domains like image classification and autonomous systems, highlighting trends and future directions.
This paper presents a summary and meta-analysis of the first three iterations of the annual International Verification of Neural Networks Competition (VNN-COMP) held in 2020, 2021, and 2022. In the VNN-COMP, participants submit software tools that analyze whether given neural networks satisfy specifications describing their input-output behavior. These neural networks and specifications cover a variety of problem classes and tasks, corresponding to safety and robustness properties in image classification, neural control, reinforcement learning, and autonomous systems. We summarize the key processes, rules, and results, present trends observed over the last three years, and provide an outlook into possible future developments.