NeurIPS Features LLNL Papers and Software

The 34th Conference on Neural Information Processing Systems (NeurIPS) features two LLNL papers advancing the reliability of deep learning for the Lab’s mission-critical applications. The most prestigious machine learning conference in the world, NeurIPS began virtually on December 6. The first paper describes a framework for understanding the effect of properties of training data on the generalization gap of machine learning (ML) algorithms—the difference between a model’s observed performance during training versus its “ground-truth” performance in the real world. The second NeurIPS paper introduces an automatic framework to obtain robustness guarantees of any deep neural network structure using the open source Linear Relaxation-based Perturbation Analysis (LiRPA) repo. Developed with colleagues at Northeastern University, China’s Tsinghua University, and UCLA, LiRPA algorithms can provide guaranteed upper and lower bounds for a neural network function with perturbed inputs.