Understanding Robustness in Teacher-Student Setting: A New Perspective

Abstract

Adversarial examples have appeared as a ubiquitous property of machine learning models where bounded adversarial perturbation could mislead the models to make arbitrarily incorrect predictions. Such examples provide a way to assess the robustness of machine learning models as well as a proxy for understanding the model training process. Extensive studies try to explain the existence of adversarial examples and provide ways to improve model robustness (e.g. adversarial training). While they mostly focus on models trained on datasets with predefined labels, we leverage the teacher-student framework and assume a teacher model, or oracle, to provide the labels for given instances. We extend Tian (2019) in the case of low-rank input data and show that student specialization (trained student neuron is highly correlated with certain teacher neuron at the same layer) still happens within the input subspace, but the teacher and student nodes could differ wildly out of the data subspace, which we conjecture leads to adversarial examples. Extensive experiments show that student specialization correlates strongly with model robustness in different scenarios, including student trained via standard training, adversarial training, confidence-calibrated adversarial training, and training with robust feature dataset. Our studies could shed light on the future exploration about adversarial examples, and enhancing model robustness via principled data augmentation.

Proceedings of the 24th International Conference on Artificial Intelligence and Statistics (AISTATS) 2021, San Diego, California, USA. PMLR: Volume 130

References in corpus (20)

• Distilling the Knowledge in a Neural Network
• Explaining and Harnessing Adversarial Examples
• Intriguing properties of neural networks
• Towards Deep Learning Models Resistant to Adversarial Attacks
• Transferability in Machine Learning: from Phenomena to Black-Box Attacks using Adversarial Samples
• Obfuscated Gradients Give a False Sense of Security: Circumventing Defenses to Adversarial Examples
• The Space of Transferable Adversarial Examples
• Robustness May Be at Odds with Accuracy
• Smooth Adversarial Training
• Testing Robustness Against Unforeseen Adversaries
• A Simple Explanation for the Existence of Adversarial Examples with Small Hamming Distance
• Exploring the Space of Black-box Attacks on Deep Neural Networks
• Unrestricted Adversarial Examples via Semantic Manipulation
• The Limitations of Adversarial Training and the Blind-Spot Attack
• On the Geometry of Adversarial Examples
• Understanding the One-Pixel Attack: Propagation Maps and Locality Analysis
• Saliency Methods for Explaining Adversarial Attacks
• Luck Matters: Understanding Training Dynamics of Deep ReLU Networks
• Understanding and Quantifying Adversarial Examples Existence in Linear Classification
• Representation Quality Of Neural Networks Links To Adversarial Attacks and Defences