LADA: Look-Ahead Data Acquisition via Augmentation for Active Learning

TL;DR

LADA is a novel framework that integrates data acquisition and augmentation in active learning, optimizing virtual data generation to improve model training efficiency and performance.

Contribution

This paper introduces LADA, a general framework that considers both data selection and augmentation simultaneously, with optimized policies like InfoMixup and InfoSTN for better active learning.

Findings

LADA significantly outperforms recent baselines on benchmark datasets.

Optimized augmentation policies improve the informativeness of virtual data.

Integrated acquisition and augmentation enhance active learning efficiency.

Abstract

Active learning effectively collects data instances for training deep learning models when the labeled dataset is limited and the annotation cost is high. Besides active learning, data augmentation is also an effective technique to enlarge the limited amount of labeled instances. However, the potential gain from virtual instances generated by data augmentation has not been considered in the acquisition process of active learning yet. Looking ahead the effect of data augmentation in the process of acquisition would select and generate the data instances that are informative for training the model. Hence, this paper proposes Look-Ahead Data Acquisition via augmentation, or LADA, to integrate data acquisition and data augmentation. LADA considers both 1) unlabeled data instance to be selected and 2) virtual data instance to be generated by data augmentation, in advance of the acquisition process. Moreover, to enhance the informativeness of the virtual data instances, LADA optimizes the data augmentation policy to maximize the predictive acquisition score, resulting in the proposal of InfoMixup and InfoSTN. As LADA is a generalizable framework, we experiment with the various combinations of acquisition and augmentation methods. The performance of LADA shows a significant improvement over the recent augmentation and acquisition baselines which were independently applied to the benchmark datasets.