The paper “Using GPUs And LLMs Can Be Satisfying for Nonlinear Real Arithmetic Problems” by Christopher Brix, Julia Walczak, Nils Lommen and Thomas Noll has been accepted for presentation at the VerifAI Workshop at ICLR 2025. The paper describes how satisfying assignments to quantifier-free NRA problems can be found using GPU acceleration, and how LLMs […]

The paper “CERTPHASH: Towards Certified Perceptual Hashing via Robust Training” by Yuchen Yang (Johns Hopkins University), Qichang Liu (Tsinghua University), Christopher Brix, Huan Zhang (UIUC), and Yinzhi Cao (Johns Hopkins University) has been accepted for publication in the 34th USENIX Security Symposium. The paper proposes a novel technique to train neural networks to perform perceptual […]

The paper “Scalable Neural Network Verification with Branch-and-bound Inferred Cutting Planes” by Duo Zhou (University of Illinois Urbana-Champaign), Christopher Brix, Grani A Hanasusanto (Illinois) and Huan Zhang (Illinois) has been accepted for the Thirty-Eighth Annual Conference on Neural Information Processing Systems (NeurIPS 2024) in Vancouver. The paper describes how the branch-and-bound process in neural network […]

The paper entitled “Provably Bounding Neural Network Preimages” by Christopher Brix, Suhas Kotha (CMU), Huan Zhang (UIUC), Zico Kolter (CMU), Krishnamurthy (Dj) Dvijotham (Google DeepMind) has been accepted for publication in the thirty-seventh Conference on Neural Information Processing Systems. The paper presents a way to compute bounds on the preimage of the output of a […]

The paper entitled “First Three Years of the International Verification of Neural Networks Competition (VNN-COMP)” by Christopher Brix, Mark Müller (ETH Zurich), Stanley Bak (Stony Brook University), Taylor T Johnson (Vanderbilt University), and Changliu Liu (Carnegie Mellon University) has been accepted for publication in the journal STTT ExPLAIn. The paper summarizes the organization and results […]