Abstract
Verification has emerged as a means to provide formal guarantees on learning-based systems incorporating neural network before using them in safety-critical applications. This paper proposes a new verification approach for deep neural networks (DNNs) with piecewise linear activation functions using reachability analysis. The core of our approach is a collection of reachability algorithms using star sets (or shortly, stars), an effective symbolic representation of high-dimensional polytopes. The star-based reachability algorithms compute the output reachable sets of a network with a given input set before using them for verification. For a neural network with piecewise linear activation functions, our approach can construct both exact and over-approximate reachable sets of the neural network. To enhance the scalability of our approach, a star set is equipped with an outer-zonotope (a zonotope over-approximation of the star set) to quickly estimate the lower and upper bounds of an input set at a specific neuron to determine if splitting occurs at that neuron. This zonotope pre-filtering step reduces significantly the number of linear programming optimization problems that must be solved in the analysis, and leads to a reduction in computation time, which enhances the scalability of the star set approach. Our reachability algorithms are implemented in a software prototype called the neural network verification tool, and can be applied to problems analyzing the robustness of machine learning methods, such as safety and robustness verification of DNNs. Our experiments show that our approach can achieve runtimes twenty to 1400 times faster than Reluplex, a satisfiability modulo theory-based approach. Our star set approach is also less conservative than other recent zonotope and abstract domain approaches.
Similar content being viewed by others
References
Akintunde ME, Kevorchian A, Lomuscio A, Pirovano E (2019) Verification of RNN-based neural agent-environment systems. In: Proceedings of the 33th AAAI conference on artificial intelligence (AAAI19). Honolulu, HI, USA. AAAI Press (to appear)
Akintunde M, Lomuscio A, Maganti L, Pirovano E (2018) Reachability analysis for neural agent-environment systems. In: Sixteenth international conference on principles of knowledge representation and reasoning
Bak S, Duggirala PS (2017) Simulation-equivalent reachability of large linear systems with inputs. In: International conference on computer aided verification. Springer, pp 401–420
Bojarski M, Del Testa D, Dworakowski D, Firner B, Flepp B, Goyal P, Jackel LD, Monfort M, Muller U, Zhang J et al (2016) End to end learning for self-driving cars. arXiv preprint arXiv:1604.07316
Bastani O, Ioannou Y, Lampropoulos L, Vytiniotis D, Nori A, Criminisi A (2016) Measuring neural net robustness with constraints. In: Advances in neural information processing systems, pp 2613–2621
Bak S, Tran H-D, Hobbs K, Johnson T (2020) Improved geometric path enumeration for verifying ReLU neural networks. In: Proceedings of the 32nd international conference on computer aided verification. Springer
Dutta S, Jha S, Sanakaranarayanan S, Tiwari A (2017) Output range analysis for deep neural networks. arXiv preprint arXiv:1709.09130
Ehlers R (2017) Formal verification of piece-wise linear feed-forward neural networks. In: International symposium on automated technology for verification and analysis. Springer, pp 269–286
Gehr T, Mirman M, Drachsler-Cohen D, Tsankov P, Chaudhuri S, Vechev M (2018) Ai 2: safety and robustness certification of neural networks with abstract interpretation. In: 2018 IEEE symposium on security and privacy (SP)
Goodfellow IJ, Shlens J, Szegedy C (2014) Explaining and harnessing adversarial examples. arXiv preprint arXiv:1412.6572
Hinton G, Deng L, Yu D, Dahl GE, Mohamed A, Jaitly N, Senior A, Vanhoucke V, Nguyen P, Sainath TN et al (2012) Deep neural networks for acoustic modeling in speech recognition: the shared views of four research groups. IEEE Signal Process Mag 29(6):82–97
Heilweil R (2020) Tesla needs to fix its deadly Autopilot problem
Huang X, Kwiatkowska M, Wang S, Wu M (2017) Safety verification of deep neural networks. In: International conference on computer aided verification. Springer, pp 3–29
Julian KD, Kochenderfer MJ, Owen MP (2018) Deep neural network compression for aircraft collision avoidance systems. arXiv preprint arXiv:1810.04240
Katz G, Barrett C, Dill DL, Julian K, Kochenderfer MJ (2017) Reluplex: an efficient smt solver for verifying deep neural networks. In: International conference on computer aided verification. Springer, pp 97–117
Kvasnica M, Grieder P, Baotić M, Morari M (2004) Multi-parametric toolbox (MPT). In: International workshop on hybrid systems: computation and control. Springer, pp 448–462
Guy K, Derek AH, Duligur I, Kyle J, Christopher L, Rachel L, Parth S, Shantanu T, Haoze W, Aleksandar Z et al (2019) The marabou framework for verification and analysis of deep neural networks. In: International conference on computer aided verification. Springer, pp 443–452
Kouvaros P, Lomuscio A (2018) Formal verification of cnn-based perception systems. arXiv preprint arXiv:1811.11373
LeCun Y (1998) The mnist database of handwritten digits. http://yann.lecun.com/exdb/mnist/
Litjens G, Kooi T, Bejnordi BE, Setio AAA, Ciompi F, Ghafoorian M, van der Laak Jeroen AWM, Van Ginneken B, Sánchez CI (2017) A survey on deep learning in medical image analysis. Med Image Anal 42:60–88
Lomuscio A, Maganti L (2017) An approach to reachability analysis for feed-forward relu neural networks. arXiv preprint arXiv:1706.07351
Liu W, Wang Z, Liu X, Zeng N, Liu Y, Alsaadi FE (2017) A survey of deep neural network architectures and their applications. Neurocomputing 234:11–26
Moosavi-Dezfooli S-M, Fawzi A, Frossard P (2016) Deepfool: a simple and accurate method to fool deep neural networks. In: Proceedings of the IEEE conference on computer vision and pattern recognition, pp 2574–2582
Muoio D (2017) The self-driving Uber in the Arizona crash was hit crossing an intersection on yellowUber crashes
Pulina L, Tacchella A (2010) An abstraction-refinement approach to verification of artificial neural networks. In: International conference on computer aided verification. Springer, pp 243–257
Singh G, Gehr T, Mirman M, Püschel M, Vechev M (2018) Fast and effective robustness certification. In: Advances in neural information processing systems, pp 10825–10836
Singh G, Gehr T, Püschel M, Vechev M (2019) An abstract domain for certifying neural networks. Proc ACM Programm Lang 3(POPL):41
Tran H-D, Bak S, Xiang W, Johnson TT (2020) Verification of deep convolutional neural networks using imagestars. In: 32nd international conference on computer-aided verification (CAV). Springer
Tran H-D, Musau P, Lopez DM, Yang X, Nguyen LV, Xiang W, Johnson TT (2019) Parallelizable reachability analysis algorithms for feed-forward neural networks. In: 7th international conference on formal methods in software engineering (FormaliSE2019),Montreal, Canada
Tran H-D, Musau P, Lopez DM, Yang X, Nguyen LV, Xiang W, Johnson TT (2019) Star-based reachability analsysis for deep neural networks. In: 23rd international symposisum on formal methods (FM’19). Springer
Tran H-D, Pal N, Musau P, Yang X, Hamilton NP, Lopez DM, Bak S, Johnson TT (2021) Robustness verification of semantic segmentation neural networks using relaxed reachability. In: Proceedings of the 33rd international conference on computeraided verification. Springer
Tran H-D, Yang X, Lopez DM, Musau P, Nguyen LV, Xiang W, Bak S, Johnson TT (2020) NNV: the neural network verification tool for deep neural networks and learning-enabled cyber-physical systems. In: 32nd international conference on computer-aided verification (CAV)
Wang S, Pei K, Whitehouse J, Yang J, Jana S (2018) Efficient formal safety analysis of neural networks. In: Advances in neural information processing systems, pp 6369–6379
Wang S, Pei K, Whitehouse J, Yang J, Jana S (2018) Formal security analysis of neural networks using symbolic intervals. arXiv preprint arXiv:1804.10829
Weng T-W, Zhang H, Chen H, Song Z, Hsieh C-J, Boning D, Dhillon IS, Daniel L (2018) Towards fast computation of certified robustness for relu networks. arXiv preprint arXiv:1804.09699
Xiang W, Musau P, Wild AA, Lopez DM, Hamilton N, Yang X, Rosenfeld JA, Johnson TT (2018) Verification for machine learning, autonomy, and neural networks survey. CoRR arXiv:1810.01989
Xiang W, Tran H-D, Johnson TT (2017) Reachable set computation and safety verification for neural networks with relu activations. arXiv preprint arXiv:1712.08163
Xiang W, Tran H-D, Johnson TT (2018) Output reachable set estimation and verification formultilayer neural networks. IEEE Trans Neural Netw Learn Syst (99):1–7
Xiang W, Tran H-D, Johnson TT (2019) Specification-guided safety verification for feedforward neural networks. AAAI Spring symposium on verification of neural networks
Zhang H, Weng T-W, Chen P-Y, Hsieh C-J, Daniel L (2018) Efficient neural network robustness certification with general activation functions. In: Advances in neural information processing systems, pp 4944–4953
Acknowledgements
The material presented in this paper is based upon work supported by the Air Force Office of Scientific Research (AFOSR) through contract numbers FA9550-18-1-0122 and FA9550-19-1-0288, the Defense Advanced Research Projects Agency (DARPA) through contract number FA8750-18-C-0089, and the National Science Foundation (NSF) through Grant Number 1910017. The U.S. government is authorized to reproduce and distribute reprints for Governmental purposes notwithstanding any copyright notation thereon. Any opinions, findings, and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of AFOSR, DARPA, or NSF.
Author information
Authors and Affiliations
Corresponding author
Additional information
Annabelle McIver, Maurice ter Beek and Cliff Jones
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
About this article
Cite this article
Tran, HD., Pal, N., Lopez, D.M. et al. Verification of piecewise deep neural networks: a star set approach with zonotope pre-filter. Form Asp Comp 33, 519–545 (2021). https://doi.org/10.1007/s00165-021-00553-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00165-021-00553-4