Greedy Policy Search: A Simple Baseline for Learnable Test-Time Augmentation

TL;DR

This paper introduces greedy policy search (GPS), a simple method for learning effective test-time augmentation policies that enhance image classification accuracy, uncertainty estimation, and robustness to domain shifts.

Contribution

The paper presents GPS, a straightforward approach to learn test-time augmentation policies that outperform fixed strategies and advanced learnable methods.

Findings

GPS-learned policies improve classification accuracy

Enhanced uncertainty estimation with GPS

Increased robustness to domain shifts

Abstract

Test-time data augmentation$-$averaging the predictions of a machine learning model across multiple augmented samples of data$-$is a widely used technique that improves the predictive performance. While many advanced learnable data augmentation techniques have emerged in recent years, they are focused on the training phase. Such techniques are not necessarily optimal for test-time augmentation and can be outperformed by a policy consisting of simple crops and flips. The primary goal of this paper is to demonstrate that test-time augmentation policies can be successfully learned too. We introduce greedy policy search (GPS), a simple but high-performing method for learning a policy of test-time augmentation. We demonstrate that augmentation policies learned with GPS achieve superior predictive performance on image classification problems, provide better in-domain uncertainty estimation, and improve the robustness to domain shift.