Reachability Analysis of Deep Neural Networks with Provable Guarantees

Ruan, Wenjie; Huang, Xiaowei; Kwiatkowska, Marta

doi:10.24963/ijcai.2018/368

Cited by 212 publications

(177 citation statements)

References 26 publications

Supporting

Mentioning

177

Contrasting

Order By: Relevance

“…A promising approach is automated verification, which aims to provide robustness guarantees for DNNs. The main relevant techniques include a layer-by-layer exhaustive search [14], methods that use constraint solvers [15], global optimisation approaches [13] and abstract interpretation [16,17] to over-approximate a DNN's behavior. Exhaustive search suffers from the state-space explosion problem, which can be alleviated by Monte Carlo tree search [2].…”

Section: Robustness Of Dnnsmentioning

confidence: 99%

Concolic testing for deep neural networks

Sun

Ruan

et al. 2018

Proceedings of the 33rd ACM/IEEE International Conference on Automated Software Engineering

Self Cite

313

249

View full text Add to dashboard Cite

Concolic testing combines program execution and symbolic analysis to explore the execution paths of a software program. This paper presents the first concolic testing approach for Deep Neural Networks (DNNs). More specifically, we formalise coverage criteria for DNNs that have been studied in the literature, and then develop a coherent method for performing concolic testing to increase test coverage. Our experimental results show the effectiveness of the concolic testing approach in both achieving high coverage and finding adversarial examples.

show abstract

Section: Robustness Of Dnnsmentioning

confidence: 99%

Concolic testing for deep neural networks

Sun

Ruan

et al. 2018

Proceedings of the 33rd ACM/IEEE International Conference on Automated Software Engineering

Self Cite

313

249

View full text Add to dashboard Cite

show abstract

“…Several works have posed the problem of certifying robustness of neural networks as a convex optimization problem. Ruan, Huang, & Kwiatkowska [85] reduce the robustness verification of a neural network to the generic reachability problem and then solve it as a convex optimization problem. Their work provides provable guarantees of upper and lower bounds, which converges to the ground truth in the limit.…”

Section: Related Workmentioning

confidence: 99%

Quantitative Verification of Neural Networks and Its Security Applications

Baluta

Shen

Shinde

et al. 2019

Proceedings of the 2019 ACM SIGSAC Conference on Computer and Communications Security

View full text Add to dashboard Cite

Neural networks are increasingly employed in safety-critical domains. This has prompted interest in verifying or certifying logically encoded properties of neural networks. Prior work has largely focused on checking existential properties, wherein the goal is to check whether there exists any input that violates a given property of interest. However, neural network training is a stochastic process, and many questions arising in their analysis require probabilistic and quantitative reasoning, i.e., estimating how many inputs satisfy a given property. To this end, our paper proposes a novel and principled framework to quantitative verification of logical properties specified over neural networks. Our framework is the first to provide PAC-style soundness guarantees, in that its quantitative estimates are within a controllable and bounded error from the true count. We instantiate our algorithmic framework by building a prototype tool called NPAQ that enables checking rich properties over binarized neural networks. We show how emerging security analyses can utilize our framework in 3 concrete point applications: quantifying robustness to adversarial inputs, efficacy of trojan attacks, and fairness/bias of given neural networks.

show abstract

“…Ruan, Huang and Kwiatkowska [44] show how to obtain the safety verification problem, the output range analysis problem and a robustness measure by instantiating the reachability problem. They present a novel algorithm based on adaptive nested optimisation to solve the reachability problem.…”

Section: Verification and Simulationmentioning

confidence: 99%

Towards formal methods and software engineering for deep learning: Security, safety and productivity for dl systems development

Hains

Jakobsson

Khmelevsky

2018

2018 Annual IEEE International Systems Conference (SysCon)

View full text Add to dashboard Cite

Deep Learning (DL) techniques are now widespread and being integrated into many important systems. Their classification and recognition abilities ensure their relevance for multiple application domains. As machine-learning that relies on training instead of algorithm programming, they offer a high degree of productivity. But they can be vulnerable to attacks and the verification of their correctness is only just emerging as a scientific and engineering possibility. This paper is a major update of a previously-published survey, attempting to cover all recent publications in this area. It also covers an even more recent trend, namely the design of domain-specific languages for producing and training neural nets.

show abstract

Reachability Analysis of Deep Neural Networks with Provable Guarantees

Cited by 212 publications

References 26 publications

Concolic testing for deep neural networks

Concolic testing for deep neural networks

Quantitative Verification of Neural Networks and Its Security Applications

Towards formal methods and software engineering for deep learning: Security, safety and productivity for dl systems development

Contact Info

Product

Resources

About