Reachable Set Computation and Safety Verification for Neural Networks with ReLU Activations

Neural networks have been widely used to solve complex real-world problems. Due to the complicate, nonlinear, non-convex nature of neural networks, formal safety guarantees for the output behaviors of neural networks will be crucial for their applications in safety-critical systems.In this paper, the output reachable set computation and safety verification problems for a class of neural networks consisting of Rectified Linear Unit (ReLU) activation functions are addressed. A layer-by-layer approach is developed to compute output reachable set. The computation is formulated in the form of a set of manipulations for a union of polyhedra, which can be efficiently applied with the aid of polyhedron computation tools. Based on the output reachable set computation results, the safety verification for a ReLU neural network can be performed by checking the intersections of unsafe regions and output reachable set described by a union of polyhedra. A numerical example of a randomly generated ReLU neural network is provided to show the effectiveness of the approach developed in this paper.

[1]  Weiming Xiang,et al.  Output Reachable Set Estimation and Verification for Multilayer Neural Networks , 2017, IEEE Transactions on Neural Networks and Learning Systems.

[2]  Ah Chung Tsoi,et al.  Face recognition: a convolutional neural-network approach , 1997, IEEE Trans. Neural Networks.

[3]  Xin Zhang,et al.  End to End Learning for Self-Driving Cars , 2016, ArXiv.

[4]  Yoshua Bengio,et al.  Deep Sparse Rectifier Neural Networks , 2011, AISTATS.

[5]  Demis Hassabis,et al.  Mastering the game of Go with deep neural networks and tree search , 2016, Nature.

[6]  Jianbin Qiu,et al.  A Combined Adaptive Neural Network and Nonlinear Model Predictive Control for Multirate Networked Industrial Process Control , 2016, IEEE Transactions on Neural Networks and Learning Systems.

[7]  Zheng-Guang Wu,et al.  Reachable Set Estimation for Markovian Jump Neural Networks With Time-Varying Delays , 2017, IEEE Transactions on Cybernetics.

[8]  Jürgen Schmidhuber,et al.  Deep learning in neural networks: An overview , 2014, Neural Networks.

[9]  Andrew L. Maas.  Rectifier Nonlinearities Improve Neural Network Acoustic Models , 2013 .

[10]  Yijing Wang,et al.  A non-ellipsoidal reachable set estimation for uncertain neural networks with time-varying delay , 2014, Commun. Nonlinear Sci. Numer. Simul..

[11]  Luca Pulina,et al.  Challenging SMT solvers to verify neural networks , 2012, AI Commun..

[12]  Min Wu,et al.  Safety Verification of Deep Neural Networks , 2016, CAV.

[13]  Peter J. Gawthrop,et al.  Neural networks for control systems - A survey , 1992, Autom..

[14]  Yan-Jun Liu,et al.  Neural Network-Based Adaptive Leader-Following Consensus Control for a Class of Nonlinear Multiagent State-Delay Systems , 2017, IEEE Transactions on Cybernetics.

[15]  Geoffrey E. Hinton,et al.  Rectified Linear Units Improve Restricted Boltzmann Machines , 2010, ICML.

[16]  Wei He,et al.  Adaptive Neural Network Control of an Uncertain Robot With Full-State Constraints , 2016, IEEE Transactions on Cybernetics.

[17]  Joan Bruna,et al.  Intriguing properties of neural networks , 2013, ICLR.

[18]  Mykel J. Kochenderfer,et al.  Reluplex: An Efficient SMT Solver for Verifying Deep Neural Networks , 2017, CAV.

[19]  Yann LeCun,et al.  What is the best multi-stage architecture for object recognition? , 2009, 2009 IEEE 12th International Conference on Computer Vision.

[20]  Manfred Morari,et al.  Multi-Parametric Toolbox 3.0 , 2013, 2013 European Control Conference (ECC).

[21]  Luca Pulina,et al.  An Abstraction-Refinement Approach to Verification of Artificial Neural Networks , 2010, CAV.

[22]  Tao Zhang,et al.  Adaptive neural network control of nonlinear systems by state and output feedback , 1999, IEEE Trans. Syst. Man Cybern. Part B.

[23]  Hieu Minh Trinh,et al.  Reachable sets bounding for generalized neural networks with interval time-varying delay and bounded disturbances , 2018, Neural Computing and Applications.

[24]  Peng Shi,et al.  Exponential Stabilization for Sampled-Data Neural-Network-Based Control Systems , 2014, IEEE Transactions on Neural Networks and Learning Systems.

[25]  Wei Xing Zheng,et al.  Synchronization and State Estimation of a Class of Hierarchical Hybrid Neural Networks With Time-Varying Delays , 2016, IEEE Transactions on Neural Networks and Learning Systems.

[26]  Wei Xing Zheng,et al.  State Estimation of Discrete-Time Switched Neural Networks With Multiple Communication Channels , 2017, IEEE Transactions on Cybernetics.

[27]  Changyin Sun,et al.  Adaptive Neural Network Control of Biped Robots , 2017, IEEE Transactions on Systems, Man, and Cybernetics: Systems.