Adversarial Self-Supervised Contrastive Learning

TL;DR

This paper introduces RoCL, a self-supervised contrastive learning method that trains neural networks to be robust against adversarial attacks without using class labels, achieving competitive robustness and transferability.

Contribution

The paper proposes a novel self-supervised adversarial training framework that does not require class labels, improving robustness and transferability of neural networks.

Findings

RoCL achieves comparable robust accuracy to supervised methods.

RoCL significantly improves robustness against black box and unseen attacks.

Joint fine-tuning enhances the model's robustness further.

Abstract

Existing adversarial learning approaches mostly use class labels to generate adversarial samples that lead to incorrect predictions, which are then used to augment the training of the model for improved robustness. While some recent works propose semi-supervised adversarial learning methods that utilize unlabeled data, they still require class labels. However, do we really need class labels at all, for adversarially robust training of deep neural networks? In this paper, we propose a novel adversarial attack for unlabeled data, which makes the model confuse the instance-level identities of the perturbed data samples. Further, we present a self-supervised contrastive learning framework to adversarially train a robust neural network without labeled data, which aims to maximize the similarity between a random augmentation of a data sample and its instance-wise adversarial perturbation. We validate our method, Robust Contrastive Learning (RoCL), on multiple benchmark datasets, on which it obtains comparable robust accuracy over state-of-the-art supervised adversarial learning methods, and significantly improved robustness against the black box and unseen types of attacks. Moreover, with further joint fine-tuning with supervised adversarial loss, RoCL obtains even higher robust accuracy over using self-supervised learning alone. Notably, RoCL also demonstrate impressive results in robust transfer learning.