First three years of the international verification of neural networks competition (VNN-COMP)

Brix, Christopher; Niklas, Müller, Mark; Bak, Stanley; Johnson, Taylor T.; Liu, Changliu

doi:10.1007/s10009-023-00703-4

Cited by 22 publications

(7 citation statements)

References 54 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This leverages open-loop neural network verification methods, which return a noisy state (the counterexample) for a fixed amount of noise. (Brix et al, 2023), where the exact noise threshold needed to cause a change in action is computed using a binary search. The minimum trajectory noise therefore depends on the tolerance of the binary search.…”

Section: Problem Statementmentioning

confidence: 99%

Provable observation noise robustness for neural network control systems

Krish,

Mata,

Bak

et al. 2024

Res. dir. Cyber phys. syst.

Self Cite

View full text Add to dashboard Cite

Neural networks are vulnerable to adversarial perturbations: slight changes to inputs that can result in unexpected outputs. In neural network control systems, these inputs are often noisy sensor readings. In such settings, natural sensor noise—or an adversary who can manipulate them—may cause the system to fail. In this paper, we introduce the first technique to provably compute the minimum magnitude of sensor noise that can cause a neural network control system to violate a safety property from a given initial state. Our algorithm constructs a tree of possible successors with increasing noise until a specification is violated. We build on open-loop neural network verification methods to determine the least amount of noise that could change actions at each step of a closed-loop execution. We prove that this method identifies the unsafe trajectory with the least noise that leads to a safety violation. We evaluate our method on four systems: the Cart Pole and LunarLander environments from OpenAI gym, an aircraft collision avoidance system based on a neural network compression of ACAS Xu, and the SafeRL Aircraft Rejoin scenario. Our analysis produces unsafe trajectories where deviations under 1% of the sensor noise range make the systems behave erroneously.

show abstract

Section: Problem Statementmentioning

confidence: 99%

Provable observation noise robustness for neural network control systems

Krish,

Mata,

Bak

et al. 2024

Res. dir. Cyber phys. syst.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Therefore, it is often impossible to simultaneously guarantee precision and efficiency at the DNN verification algorithm level. Even many state of the art verification tools [26] still have high running time while ensuring precision.…”

Section: Formal Verification and Parallelizationmentioning

confidence: 99%

“…Current verification techniques can be divided into CPU-based and GPU-based. Most tools support and adopt CPU verification mode [26], which can be parallelized in a multi-core environment and involves plenty of calculations and logical judgment operations. In this context, we can classify verification queries as CPU-intensive tasks that are processed asynchronously by multicore CPUs.…”

Section: Parallel Frameworkmentioning

confidence: 99%

“…Although formal verification is an important technique for the safety evaluation of deep learning models, the inefficiency caused by the exact encoding of complex structures of the models is inevitable, which has become the main limitation. 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].…”

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

“…The Verification of Neural Networks Competition (VNN-COMP) aims to evaluate these solvers on a suite of benchmarks consisting of neural networks trained on practical problems. After the first three iterations of this competition, its organizers give an overview over its results and the current status quo in neural network verification in their paper "First Three Years of the International Verification of Neural Networks Competition (VNN-COMP)" [8]. They discuss the goals and results of this competition and the possible next steps to further develop the competition setting.…”

Section: Verification Of Neural Network Competitionmentioning

confidence: 99%

Explanation Paradigms Leveraging Analytic Intuition (ExPLAIn)

Jansen,

Nolte,

Steffen

2023

Int J Softw Tools Technol Transfer

View full text Add to dashboard Cite

In this paper, we present the envisioned style and scope of the new topic “Explanation Paradigms Leveraging Analytic Intuition” (ExPLAIn) with the International Journal on Software Tools for Technology Transfer (STTT). Intention behind this new topic is to (1) explicitly address all aspects and issues that arise when trying to, if possible, reveal and then confirm hidden properties of black-box systems, or (2) to enforce vital properties by embedding them into appropriate system contexts. Machine-learned systems, such as Deep Neural Networks, are particularly challenging black-box systems, and there is a wealth of formal methods for analysis and verification waiting to be adapted and applied. The selection of papers of this first Special Section of ExPLAIn, most of which were co-authored by editorial board members, is an illustrative example of the style and scope envisioned: In addition to methodological papers on verification, explanation, and their scalability, case studies, tool papers, literature reviews, and position papers are also welcome.

show abstract

First three years of the international verification of neural networks competition (VNN-COMP)

Cited by 22 publications

References 54 publications

Provable observation noise robustness for neural network control systems

Provable observation noise robustness for neural network control systems

A Parallel Optimization Method for Robustness Verification of Deep Neural Networks

Explanation Paradigms Leveraging Analytic Intuition (ExPLAIn)

Contact Info

Product

Resources

About