Cargando…

NNV: The Neural Network Verification Tool for Deep Neural Networks and Learning-Enabled Cyber-Physical Systems

This paper presents the Neural Network Verification (NNV) software tool, a set-based verification framework for deep neural networks (DNNs) and learning-enabled cyber-physical systems (CPS). The crux of NNV is a collection of reachability algorithms that make use of a variety of set representations,...

Descripción completa

Detalles Bibliográficos
Autores principales: Tran, Hoang-Dung, Yang, Xiaodong, Manzanas Lopez, Diego, Musau, Patrick, Nguyen, Luan Viet, Xiang, Weiming, Bak, Stanley, Johnson, Taylor T.
Formato: Online Artículo Texto
Lenguaje:English
Publicado: 2020
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7363192/
http://dx.doi.org/10.1007/978-3-030-53288-8_1
_version_ 1783559621181440000
author Tran, Hoang-Dung
Yang, Xiaodong
Manzanas Lopez, Diego
Musau, Patrick
Nguyen, Luan Viet
Xiang, Weiming
Bak, Stanley
Johnson, Taylor T.
author_facet Tran, Hoang-Dung
Yang, Xiaodong
Manzanas Lopez, Diego
Musau, Patrick
Nguyen, Luan Viet
Xiang, Weiming
Bak, Stanley
Johnson, Taylor T.
author_sort Tran, Hoang-Dung
collection PubMed
description This paper presents the Neural Network Verification (NNV) software tool, a set-based verification framework for deep neural networks (DNNs) and learning-enabled cyber-physical systems (CPS). The crux of NNV is a collection of reachability algorithms that make use of a variety of set representations, such as polyhedra, star sets, zonotopes, and abstract-domain representations. NNV supports both exact (sound and complete) and over-approximate (sound) reachability algorithms for verifying safety and robustness properties of feed-forward neural networks (FFNNs) with various activation functions. For learning-enabled CPS, such as closed-loop control systems incorporating neural networks, NNV provides exact and over-approximate reachability analysis schemes for linear plant models and FFNN controllers with piecewise-linear activation functions, such as ReLUs. For similar neural network control systems (NNCS) that instead have nonlinear plant models, NNV supports over-approximate analysis by combining the star set analysis used for FFNN controllers with zonotope-based analysis for nonlinear plant dynamics building on CORA. We evaluate NNV using two real-world case studies: the first is safety verification of ACAS Xu networks, and the second deals with the safety verification of a deep learning-based adaptive cruise control system.
format Online
Article
Text
id pubmed-7363192
institution National Center for Biotechnology Information
language English
publishDate 2020
record_format MEDLINE/PubMed
spelling pubmed-73631922020-07-16 NNV: The Neural Network Verification Tool for Deep Neural Networks and Learning-Enabled Cyber-Physical Systems Tran, Hoang-Dung Yang, Xiaodong Manzanas Lopez, Diego Musau, Patrick Nguyen, Luan Viet Xiang, Weiming Bak, Stanley Johnson, Taylor T. Computer Aided Verification Article This paper presents the Neural Network Verification (NNV) software tool, a set-based verification framework for deep neural networks (DNNs) and learning-enabled cyber-physical systems (CPS). The crux of NNV is a collection of reachability algorithms that make use of a variety of set representations, such as polyhedra, star sets, zonotopes, and abstract-domain representations. NNV supports both exact (sound and complete) and over-approximate (sound) reachability algorithms for verifying safety and robustness properties of feed-forward neural networks (FFNNs) with various activation functions. For learning-enabled CPS, such as closed-loop control systems incorporating neural networks, NNV provides exact and over-approximate reachability analysis schemes for linear plant models and FFNN controllers with piecewise-linear activation functions, such as ReLUs. For similar neural network control systems (NNCS) that instead have nonlinear plant models, NNV supports over-approximate analysis by combining the star set analysis used for FFNN controllers with zonotope-based analysis for nonlinear plant dynamics building on CORA. We evaluate NNV using two real-world case studies: the first is safety verification of ACAS Xu networks, and the second deals with the safety verification of a deep learning-based adaptive cruise control system. 2020-06-13 /pmc/articles/PMC7363192/ http://dx.doi.org/10.1007/978-3-030-53288-8_1 Text en © The Author(s) 2020 Open Access This chapter is licensed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made. The images or other third party material in this chapter are included in the chapter's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the chapter's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder.
spellingShingle Article
Tran, Hoang-Dung
Yang, Xiaodong
Manzanas Lopez, Diego
Musau, Patrick
Nguyen, Luan Viet
Xiang, Weiming
Bak, Stanley
Johnson, Taylor T.
NNV: The Neural Network Verification Tool for Deep Neural Networks and Learning-Enabled Cyber-Physical Systems
title NNV: The Neural Network Verification Tool for Deep Neural Networks and Learning-Enabled Cyber-Physical Systems
title_full NNV: The Neural Network Verification Tool for Deep Neural Networks and Learning-Enabled Cyber-Physical Systems
title_fullStr NNV: The Neural Network Verification Tool for Deep Neural Networks and Learning-Enabled Cyber-Physical Systems
title_full_unstemmed NNV: The Neural Network Verification Tool for Deep Neural Networks and Learning-Enabled Cyber-Physical Systems
title_short NNV: The Neural Network Verification Tool for Deep Neural Networks and Learning-Enabled Cyber-Physical Systems
title_sort nnv: the neural network verification tool for deep neural networks and learning-enabled cyber-physical systems
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7363192/
http://dx.doi.org/10.1007/978-3-030-53288-8_1
work_keys_str_mv AT tranhoangdung nnvtheneuralnetworkverificationtoolfordeepneuralnetworksandlearningenabledcyberphysicalsystems
AT yangxiaodong nnvtheneuralnetworkverificationtoolfordeepneuralnetworksandlearningenabledcyberphysicalsystems
AT manzanaslopezdiego nnvtheneuralnetworkverificationtoolfordeepneuralnetworksandlearningenabledcyberphysicalsystems
AT musaupatrick nnvtheneuralnetworkverificationtoolfordeepneuralnetworksandlearningenabledcyberphysicalsystems
AT nguyenluanviet nnvtheneuralnetworkverificationtoolfordeepneuralnetworksandlearningenabledcyberphysicalsystems
AT xiangweiming nnvtheneuralnetworkverificationtoolfordeepneuralnetworksandlearningenabledcyberphysicalsystems
AT bakstanley nnvtheneuralnetworkverificationtoolfordeepneuralnetworksandlearningenabledcyberphysicalsystems
AT johnsontaylort nnvtheneuralnetworkverificationtoolfordeepneuralnetworksandlearningenabledcyberphysicalsystems