Understanding Robust Learning through the Lens of Representation Similarities

TL;DR

This paper investigates how representations learned by robust neural networks differ from standard ones, revealing key properties and challenges that inform future robust model design.

Contribution

It introduces a comprehensive analysis of robust versus non-robust representations using similarity metrics across multiple datasets, uncovering novel properties of robust networks.

Findings

Robust networks lack representation specialization.

Disappearance of block structure in robust representations.

Overfitting mainly affects deeper layers during robust training.

Abstract

Representation learning, i.e. the generation of representations useful for downstream applications, is a task of fundamental importance that underlies much of the success of deep neural networks (DNNs). Recently, robustness to adversarial examples has emerged as a desirable property for DNNs, spurring the development of robust training methods that account for adversarial examples. In this paper, we aim to understand how the properties of representations learned by robust training differ from those obtained from standard, non-robust training. This is critical to diagnosing numerous salient pitfalls in robust networks, such as, degradation of performance on benign inputs, poor generalization of robustness, and increase in over-fitting. We utilize a powerful set of tools known as representation similarity metrics, across three vision datasets, to obtain layer-wise comparisons between robust and non-robust DNNs with different training procedures, architectural parameters and adversarial constraints. Our experiments highlight hitherto unseen properties of robust representations that we posit underlie the behavioral differences of robust networks. We discover a lack of specialization in robust networks' representations along with a disappearance of `block structure'. We also find overfitting during robust training largely impacts deeper layers. These, along with other findings, suggest ways forward for the design and training of better robust networks.