Neuron Similarity-Based Neural Network Verification via Abstraction and Refinement

Authors: Yuehao Liu, Yansong Dong, Liang Zhao, Wensheng Wang, Cong Tian

IJCAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Experimental results demonstrate that ARVerifier significantly reduces network size and yields verification time reductions by 11.61%, 18.70%, and 12.20% compared to α, β-CROWN, Verinet, and Marabou, respectively.
Researcher Affiliation Collaboration 1School of Computer Science and Technology, Xidian University 2Beijing Sunwise Information Technology Ltd EMAIL, EMAIL, EMAIL, EMAIL
Pseudocode Yes Algorithm 1 specifies the general neural network verification method based on CEGAR. Given a verification problem N, P, Q , the algorithm returns SAFE if the safety property is satisfied, or UNSAFE along with a counterexample cex.
Open Source Code No The paper states, "We have implemented this method as a tool named ARVerifier and integrated it with three state-of-the-art verification tools for evaluation on ACAS Xu and MNIST benchmarks." While it mentions implementation as a tool, there is no explicit statement about making the source code publicly available or providing a link to a repository.
Open Datasets Yes The neural networks considered in the experiments are from the fully connected benchmarks used in the international verification of neural networks competition (VNN-COMP) [Brix et al., 2024; Brix et al., 2023], specifically the ACAS Xu and MNIST datasets.
Dataset Splits No The paper describes how properties were tested on the ACAS Xu and MNIST datasets (e.g., "testing 4 properties across all 45 networks and 6 on a single network" for ACAS Xu, and "testing 25 images with l perturbations of 0.02 and 0.05" for MNIST). However, it does not specify any training/test/validation splits for the datasets themselves, as the paper focuses on verification of pre-existing DNNs, not their training.
Hardware Specification Yes Experiments are performed on a 64-bit Ubuntu 18.04 platform equipped with 64 GB of RAM and an Intel i77700 quad-core processor. ... α, β-CROWN is executed with an additional NVIDIA TITAN RTX GPU with 24 GB memory.
Software Dependencies Yes All verification tools are implemented in Python 3.8. For the tools involving MILP solving, α, β-CROWN and Marabou use Gurobi 9.1, while Verinet uses Xpress 9.0 as the backend solver.
Experiment Setup Yes Besides, we impose a 30-minute timeout for each verification problem, while maintaining default values for other parameters.