Notes on Margin Training and Margin p-Values for Deep Neural Network Classifiers

2019/10/15 by George Kesidis, Kesidis, George, David J. Miller +3
Computer Science · #Adversarial Robustness in Machine Learning #Anomaly Detection Techniques and Applications #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning and Algorithms

paper · pdf · doi:10.48550/arxiv.1910.08032

openalex publication_date 2019/10/15 · openalex created_date 2019/12/13 · openalex updated_date 2026/07/28

Abstract

We provide a new local class-purity theorem for Lipschitz continuous DNN classifiers. In addition, we discuss how to achieve classification margin for training samples. Finally, we describe how to compute margin p-values for test samples.

Citations

• Parseval Networks: Improving Robustness to Adversarial Examples
• Lipschitz-Margin Training: Scalable Certification of Perturbation Invariance for Deep Neural Networks
• Robustness May Be at Odds with Accuracy

Related

• Distribution of Classification Margins: Are All Data Equal?
• Certified Robustness via Dynamic Margin Maximization and Improved Lipschitz Regularization
• Recent Advances in Large Margin Learning
• Margin-Based Feed-Forward Neural Network Classifiers
• Lipschitz-Margin Training: Scalable Certification of Perturbation Invariance for Deep Neural Networks