UniformAugment: A Search-free Probabilistic Data Augmentation Approach

TL;DR

UniformAugment introduces a search-free, probabilistic data augmentation method that efficiently improves model performance by uniformly sampling augmentation transformations, eliminating the need for computationally expensive search phases.

Contribution

The paper proposes UniformAugment, a novel data augmentation approach that removes the search process by assuming distribution invariance, maintaining effectiveness while greatly reducing computational costs.

Findings

UniformAugment achieves comparable accuracy to search-based methods.

It significantly reduces computational overhead in data augmentation.

The approach is effective across standard datasets and models.

Abstract

Augmenting training datasets has been shown to improve the learning effectiveness for several computer vision tasks. A good augmentation produces an augmented dataset that adds variability while retaining the statistical properties of the original dataset. Some techniques, such as AutoAugment and Fast AutoAugment, have introduced a search phase to find a set of suitable augmentation policies for a given model and dataset. This comes at the cost of great computational overhead, adding up to several thousand GPU hours. More recently RandAugment was proposed to substantially speedup the search phase by approximating the search space by a couple of hyperparameters, but still incurring non-negligible cost for tuning those. In this paper we show that, under the assumption that the augmentation space is approximately distribution invariant, a uniform sampling over the continuous space of augmentation transformations is sufficient to train highly effective models. Based on that result we propose UniformAugment, an automated data augmentation approach that completely avoids a search phase. In addition to discussing the theoretical underpinning supporting our approach, we also use the standard datasets, as well as established models for image classification, to show that UniformAugment's effectiveness is comparable to the aforementioned methods, while still being highly efficient by virtue of not requiring any search.