Adaptive Adversarial Logits Pairing

TL;DR

This paper introduces Adaptive Adversarial Logits Pairing (AALP), a novel training method that improves adversarial robustness by adaptively focusing on key features and balancing training objectives, outperforming previous approaches.

Contribution

AALP modifies adversarial training with adaptive feature selection and sample weighting, enhancing robustness against adversarial attacks.

Findings

AALP achieves superior defense performance on multiple datasets.

Adaptive modules improve focus on high-contribution features.

Balanced training reduces overemphasis on logits pairing loss.

Abstract

Adversarial examples provide an opportunity as well as impose a challenge for understanding image classification systems. Based on the analysis of the adversarial training solution Adversarial Logits Pairing (ALP), we observed in this work that: (1) The inference of adversarially robust model tends to rely on fewer high-contribution features compared with vulnerable ones. (2) The training target of ALP doesn't fit well to a noticeable part of samples, where the logits pairing loss is overemphasized and obstructs minimizing the classification loss. Motivated by these observations, we design an Adaptive Adversarial Logits Pairing (AALP) solution by modifying the training process and training target of ALP. Specifically, AALP consists of an adaptive feature optimization module with Guided Dropout to systematically pursue fewer high-contribution features, and an adaptive sample weighting module by setting sample-specific training weights to balance between logits pairing loss and classification loss. The proposed AALP solution demonstrates superior defense performance on multiple datasets with extensive experiments.