Abstract
We introduce ReachNN*, a tool for reachability analysis of neural-network controlled systems (NNCSs). The theoretical foundation of ReachNN* is the use of Bernstein polynomials to approximate any Lipschitz-continuous neural-network controller with different types of activation functions, with provable approximation error bounds. In addition, the sampling-based error bound estimation in ReachNN* is amenable to GPU-based parallel computing. For further improvement in runtime and error bound estimation, ReachNN* also features optional controller re-synthesis via a technique called verification-aware knowledge distillation (KD) to reduce the Lipschitz constant of the neural-network controller. Experiment results across a set of benchmarks show \(7\times \) to \(422\times \) efficiency improvement over the previous prototype. Moreover, KD enables proof of reachability of NNCSs whose verification results were previously unknown due to large overapproximation errors. An open-source implementation of ReachNN* is available at https://github.com/JmfanBU/ReachNNStar.git.
J. Fan and C. Huang contributed equally.
We acknowledge the support from NSF grants 1646497, 1834701, 1834324, 1839511, 1724341, ONR grant N00014-19-1-2496, and the US Air Force Research Laboratory (AFRL) under contract number FA8650-16-C-2642.
This is a preview of subscription content, log in via an institution.
Buying options
Tax calculation will be finalised at checkout
Purchases are for personal use only
Learn about institutional subscriptionsReferences
Abadi, M., et al.: Tensorflow: a system for large-scale machine learning. In: OSDI, pp. 265–283 (2016)
Bogomolov, S., Forets, M., Frehse, G., Potomkin, K., Schilling, C.: Juliareach: a toolbox for set-based reachability. In: HSCC, pp. 39–44 (2019)
Chen, S., et al.: Approximating explicit model predictive control using constrained neural networks. In: ACC, pp. 1520–1527. IEEE (2018)
Chen, X., Ábrahám, E., Sankaranarayanan, S.: Flow*: an analyzer for non-linear hybrid systems. In: Sharygina, N., Veith, H. (eds.) CAV 2013. LNCS, vol. 8044, pp. 258–263. Springer, Heidelberg (2013). https://doi.org/10.1007/978-3-642-39799-8_18
Dutta, S., Chen, X., Sankaranarayanan, S.: Reachability analysis for neural feedback systems using regressive polynomial rule inference. In: HSCC, pp. 157–168 (2019)
Fan, J., Huang, C., Li, W., Chen, X., Zhu, Q.: Towards verification-aware knowledge distillation for neural-network controlled systems. In: ICCAD. IEEE (2019)
Finn, C., Yu, T., Zhang, T., Abbeel, P., Levine, S.: One-shot visual imitation learning via meta-learning. In: Conference on Robot Learning, pp. 357–368 (2017)
Gallestey, E., Hokayem, P.: Lecture notes in nonlinear systems and control (2019)
Hertneck, M., Köhler, J., Trimpe, S., Allgöwer, F.: Learning an approximate model predictive controller with guarantees. IEEE Control Syst. Lett. 2(3), 543–548 (2018)
Hinton, G.E., Vinyals, O., Dean, J.: Distilling the knowledge in a neural network. CoRR abs/1503.02531 (2015)
Huang, C., Fan, J., Li, W., Chen, X., Zhu, Q.: REACHNN: reachability analysis of neural-network controlled systems. TECS 18(5s), 1–22 (2019)
Ivanov, R., Weimer, J., Alur, R., Pappas, G.J., Lee, I.: Verisig: verifying safety properties of hybrid systems with neural network controllers. In: HSCC, pp. 169–178 (2019)
Lillicrap, T.P., et al.: Continuous control with deep reinforcement learning. In: International Conference on Learning Representation (2016)
Pan, Y., et al.: Agile autonomous driving using end-to-end deep imitation learning. In: RSS (2018)
Tran, H.D., Cai, F., Diego, M.L., Musau, P., Johnson, T.T., Koutsoukos, X.: Safety verification of cyber-physical systems with reinforcement learning control. TECS 18(5s), 1–22 (2019)
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2020 Springer Nature Switzerland AG
About this paper
Cite this paper
Fan, J., Huang, C., Chen, X., Li, W., Zhu, Q. (2020). ReachNN*: A Tool for Reachability Analysis of Neural-Network Controlled Systems. In: Hung, D.V., Sokolsky, O. (eds) Automated Technology for Verification and Analysis. ATVA 2020. Lecture Notes in Computer Science(), vol 12302. Springer, Cham. https://doi.org/10.1007/978-3-030-59152-6_30
Download citation
DOI: https://doi.org/10.1007/978-3-030-59152-6_30
Published:
Publisher Name: Springer, Cham
Print ISBN: 978-3-030-59151-9
Online ISBN: 978-3-030-59152-6
eBook Packages: Computer ScienceComputer Science (R0)