Artifact for article: PRIMA: General and Precise Neural Network Certification via Scalable Convex Hull Approximations

Niklas, Müller, Mark; Makarchuk, Gleb; Singh, Gagandeep; Püschel, Markus; Vechev, Martin

doi:10.1145/3462308

Cited by 9 publications

(4 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Additionally, the local robustness verification of a single input point is often lack of convincing and representative, and the computational cost of global robustness verification is too expensive. Thus, we can integrate the parallel verification with the GPU mode [30] to explore some more complex security properties.…”

Section: Discussionmentioning

confidence: 99%

“…Even state-of-the-art tools [22][23][24][25] in the neural network verification competition [26] face the lack of performance and scalability when handling complex networks and large quantities of tasks. In light of this, a feasible optimization scheme is parallelization [27][28][29][30]. Computational complexity is a long-standing problem in verification technologies, distributing tasks to multiple workers in parallel for simultaneous computation can improve the solving performance and the number of applicable input tasks.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

A Parallel Optimization Method for Robustness Verification of Deep Neural Networks

Lin,

Zhou,

Nan

et al. 2024

Mathematics

View full text Add to dashboard Cite

Deep neural networks (DNNs) have gained considerable attention for their expressive capabilities, but unfortunately they have serious robustness risks. Formal verification is an important technique to ensure network reliability. However, current verification techniques are unsatisfactory in time performance, which hinders the practical applications. To address this issue, we propose an efficient optimization method based on parallel acceleration with more computing resources. The method involves the speedup configuration of a partition-based verification aligned with the structures and robustness formal specifications of DNNs. A parallel verification framework is designed specifically for neural network verification systems, which integrates various auxiliary modules and accommodates diverse verification modes. The efficient parallel scheduling of verification queries within the framework enhances resource utilization and enables the system to process a substantial volume of verification tasks. We conduct extensive experiments on multiple commonly used verification benchmarks to demonstrate the rationality and effectiveness of the proposed method. The results show that higher efficiency is achieved after parallel optimization integration.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

A Parallel Optimization Method for Robustness Verification of Deep Neural Networks

Lin,

Zhou,

Nan

et al. 2024

Mathematics

View full text Add to dashboard Cite

show abstract

“…Thus, relating system-level specifications to notions of input distance is sometimes complex. For a given distance definition, formal verification methods attempt to formally prove certain properties of DL models, including robustness (Katz et al, 2017;Mirman et al, 2018;Müller et al, 2021;Wang et al, 2018). They allow a user to build a model tolerant to a certain distance between inputs.…”

Section: Robustness and Reliabilitymentioning

confidence: 99%

A critical review on the state-of-the-art and future prospects of machine learning for Earth observation operations

Miralles

Thangavel

Scannapieco

et al. 2023

Advances in Space Research

View full text Add to dashboard Cite

“…return UNSAT Before invoking solveConv to solve φ, it is common to first call an efficient bound-tightening procedure (tightenBounds) to prune the search space or even derive UNSAT preemptively. This tightenBounds procedure can be instantiated in various ways, including with analyses based on LiPRA [74,76,58,70], kReLU [56], or PRIMA [49]. In addition to the dedicated bound-tightening pass, some convex procedures (e.g., Simplex) also naturally lend themselves to bound inference during their executions [38,35].…”

Section: Preliminariesmentioning

confidence: 99%

Efficient Neural Network Analysis with Sum-of-Infeasibilities

Zeljić

Katz

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Inspired by sum-of-infeasibilities methods in convex optimization, we propose a novel procedure for analyzing verification queries on neural networks with piecewise-linear activation functions. Given a convex relaxation which over-approximates the non-convex activation functions, we encode the violations of activation functions as a cost function and optimize it with respect to the convex relaxation. The cost function, referred to as the Sum-of-Infeasibilities (SoI), is designed so that its minimum is zero and achieved only if all the activation functions are satisfied. We propose a stochastic procedure, , to efficiently minimize the SoI. An extension to a canonical case-analysis-based complete search procedure can be achieved by replacing the convex procedure executed at each search state with . Extending the complete search with achieves multiple simultaneous goals: 1) it guides the search towards a counter-example; 2) it enables more informed branching decisions; and 3) it creates additional opportunities for bound derivation. An extensive evaluation across different benchmarks and solvers demonstrates the benefit of the proposed techniques. In particular, we demonstrate that SoI significantly improves the performance of an existing complete search procedure. Moreover, the SoI-based implementation outperforms other state-of-the-art complete verifiers. We also show that our technique can efficiently improve upon the perturbation bound derived by a recent adversarial attack algorithm.

show abstract

Artifact for article: PRIMA: General and Precise Neural Network Certification via Scalable Convex Hull Approximations

Cited by 9 publications

References 0 publications

A Parallel Optimization Method for Robustness Verification of Deep Neural Networks

A Parallel Optimization Method for Robustness Verification of Deep Neural Networks

A critical review on the state-of-the-art and future prospects of machine learning for Earth observation operations

Efficient Neural Network Analysis with Sum-of-Infeasibilities

Contact Info

Product

Resources

About