Layer-wise Sustainability Analysis

Layer-wise Regularized Adversarial Training Using Layers Sustainability Analysis Framework

Authors

Mohammad Khalooei, Mohammad Mehdi Homayounpour, Maryam Amirmazlaghani

Abstract

• The layer sustainability analysis (LSA) framework is introduced to evaluate the behavior of layer-level representations of DNNs in dealing with network input perturbations using Lipschitz theoretical concepts.
• A layer-wise regularized adversarial training (AT-LR) approach significantly improves the generalization and robustness of different deep neural network architectures for significant perturbations while reducing layer-level vulnerabilities.
• AT-LR loss landscapes for each LSA MVL proposal can interpret layer importance for different layers, which is an intriguing aspect.

Links

• ArXiv Paper
• GitHub Repository
• ScienceDirect Article