General cutting planes for bound-propagation-based neural network verification

H Zhang, S Wang, K Xu, L Li, B Li… - Advances in neural …, 2022 - proceedings.neurips.cc
Bound propagation methods, when combined with branch and bound, are among the most
effective methods to formally verify properties of deep neural networks such as correctness …

Beta-crown: Efficient bound propagation with per-neuron split constraints for neural network robustness verification

S Wang, H Zhang, K Xu, X Lin, S Jana… - Advances in …, 2021 - proceedings.neurips.cc
Bound propagation based incomplete neural network verifiers such as CROWN are very
efficient and can significantly accelerate branch-and-bound (BaB) based complete …

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

C Brix, MN Müller, S Bak, TT Johnson, C Liu - International Journal on …, 2023 - Springer
This paper presents a summary and meta-analysis of the first three iterations of the annual
International Verification of Neural Networks Competition (VNN-COMP), held in 2020, 2021 …

Sok: Certified robustness for deep neural networks

L Li, T **e, B Li - 2023 IEEE symposium on security and privacy …, 2023 - ieeexplore.ieee.org
Great advances in deep neural networks (DNNs) have led to state-of-the-art performance on
a wide range of tasks. However, recent studies have shown that DNNs are vulnerable to …

The second international verification of neural networks competition (vnn-comp 2021): Summary and results

S Bak, C Liu, T Johnson - arxiv preprint arxiv:2109.00498, 2021 - arxiv.org
This report summarizes the second International Verification of Neural Networks
Competition (VNN-COMP 2021), held as a part of the 4th Workshop on Formal Methods for …

Training certifiably robust neural networks with efficient local lipschitz bounds

Y Huang, H Zhang, Y Shi, JZ Kolter… - Advances in Neural …, 2021 - proceedings.neurips.cc
Certified robustness is a desirable property for deep neural networks in safety-critical
applications, and popular training algorithms can certify robustness of a neural network by …

Efficiently computing local lipschitz constants of neural networks via bound propagation

Z Shi, Y Wang, H Zhang, JZ Kolter… - Advances in Neural …, 2022 - proceedings.neurips.cc
Lipschitz constants are connected to many properties of neural networks, such as
robustness, fairness, and generalization. Existing methods for computing Lipschitz constants …

PRIMA: general and precise neural network certification via scalable convex hull approximations

MN Müller, G Makarchuk, G Singh, M Püschel… - Proceedings of the …, 2022 - dl.acm.org
Formal verification of neural networks is critical for their safe adoption in real-world
applications. However, designing a precise and scalable verifier which can handle different …

The third international verification of neural networks competition (VNN-COMP 2022): Summary and results

MN Müller, C Brix, S Bak, C Liu, TT Johnson - arxiv preprint arxiv …, 2022 - arxiv.org
This report summarizes the 3rd International Verification of Neural Networks Competition
(VNN-COMP 2022), held as a part of the 5th Workshop on Formal Methods for ML-Enabled …

Complete verification via multi-neuron relaxation guided branch-and-bound

C Ferrari, MN Muller, N Jovanovic… - arxiv preprint arxiv …, 2022 - arxiv.org
State-of-the-art neural network verifiers are fundamentally based on one of two paradigms:
either encoding the whole verification problem via tight multi-neuron convex relaxations or …