Informative Dropout for Robust Representation Learning: A Shape-bias Perspective

TL;DR

This paper introduces InfoDrop, a lightweight, model-agnostic method inspired by human vision, that reduces CNN texture bias to improve robustness across various scenarios like domain shifts, adversarial attacks, and corruptions.

Contribution

It proposes a novel, interpretable dropout-based technique to decrease texture bias in CNNs, enhancing their robustness and trustworthiness in diverse conditions.

Findings

Enhanced robustness in CNNs across multiple scenarios

Reduced texture bias improves interpretability

Unified approach benefits various robustness types

Abstract

Convolutional Neural Networks (CNNs) are known to rely more on local texture rather than global shape when making decisions. Recent work also indicates a close relationship between CNN's texture-bias and its robustness against distribution shift, adversarial perturbation, random corruption, etc. In this work, we attempt at improving various kinds of robustness universally by alleviating CNN's texture bias. With inspiration from the human visual system, we propose a light-weight model-agnostic method, namely Informative Dropout (InfoDrop), to improve interpretability and reduce texture bias. Specifically, we discriminate texture from shape based on local self-information in an image, and adopt a Dropout-like algorithm to decorrelate the model output from the local texture. Through extensive experiments, we observe enhanced robustness under various scenarios (domain generalization, few-shot classification, image corruption, and adversarial perturbation). To the best of our knowledge, this work is one of the earliest attempts to improve different kinds of robustness in a unified model, shedding new light on the relationship between shape-bias and robustness, also on new approaches to trustworthy machine learning algorithms. Code is available at https://github.com/bfshi/InfoDrop.