LGCVNov 2, 2022

Adversarial Auto-Augment with Label Preservation: A Representation Learning Principle Guided Approach

UW
arXiv:2211.00824v114 citationsh-index: 42Has Code
Originality Highly original
AI Analysis

This work addresses the need for autonomous data augmentation in machine learning, particularly in domains like medical imaging where domain knowledge is scarce, offering a novel method that integrates into various tasks without extra generative models.

The paper tackles the problem of data augmentation without prior domain knowledge by deriving an objective from a representation learning principle to create hard positive examples that preserve labels, and shows consistent improvements in supervised, semi-supervised, and noisy-label learning tasks, including on medical images where existing techniques perform poorly.

Data augmentation is a critical contributing factor to the success of deep learning but heavily relies on prior domain knowledge which is not always available. Recent works on automatic data augmentation learn a policy to form a sequence of augmentation operations, which are still pre-defined and restricted to limited options. In this paper, we show that a prior-free autonomous data augmentation's objective can be derived from a representation learning principle that aims to preserve the minimum sufficient information of the labels. Given an example, the objective aims at creating a distant "hard positive example" as the augmentation, while still preserving the original label. We then propose a practical surrogate to the objective that can be optimized efficiently and integrated seamlessly into existing methods for a broad class of machine learning tasks, e.g., supervised, semi-supervised, and noisy-label learning. Unlike previous works, our method does not require training an extra generative model but instead leverages the intermediate layer representations of the end-task model for generating data augmentations. In experiments, we show that our method consistently brings non-trivial improvements to the three aforementioned learning tasks from both efficiency and final performance, either or not combined with strong pre-defined augmentations, e.g., on medical images when domain knowledge is unavailable and the existing augmentation techniques perform poorly. Code is available at: https://github.com/kai-wen-yang/LPA3}{https://github.com/kai-wen-yang/LPA3.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes