ReachNN*: Reachability Analysis Tool of Neural Network Controlled Systems (NNCSs)
- ReachNN* uses Bernstein polynomials to approximate neural networks with general types of activation functions. It also offers a feature to automatically retrain a verification-friendly network.
- Github link: https://github.com/JmfanBU/ReachNNStar
- Related publications:
- Jiameng Fan, Chao Huang, Xin Chen, Wenchao Li and Qi Zhu. ReachNN*: A Tool for Reachability Analysis ofNeural-Network Controlled Systems. The 18th International Symposium on Automated Technology for Verification and Analysis (ATVA), October 2020 (to appear).
- C.Huang, J.Fan, W.Li, X.Chen, and Q.Zhu. ReachNN: Reachability Analysis of Neural-Network Controlled Systems. ACM Transactions on Embedded Computing Systems (TECS), 18:1–22, October 2019.
- J.Fan, C.Huang, W.Li, X.Chen, and Q.Zhu. Towards Verification-Aware Knowledge Distillation for Neural-Network Controlled Systems. International Conference on Computer Aided Design (ICCAD), November 2019.