Learning Data Manipulation for Augmentation and Weighting

TL;DR

This paper introduces a unified gradient-based method for learning various data manipulation techniques, such as augmentation and weighting, by leveraging reinforcement learning principles to improve classification tasks especially in low-data and imbalanced scenarios.

Contribution

It proposes a novel, flexible approach that learns different data manipulation schemes using a single gradient-based algorithm, connecting supervised learning with reinforcement learning.

Findings

Significant improvements in image and text classification accuracy.

Effective data augmentation via learned text transformation networks.

Enhanced performance in low-data and class-imbalance settings.

Abstract

Manipulating data, such as weighting data examples or augmenting with new instances, has been increasingly used to improve model training. Previous work has studied various rule- or learning-based approaches designed for specific types of data manipulation. In this work, we propose a new method that supports learning different manipulation schemes with the same gradient-based algorithm. Our approach builds upon a recent connection of supervised learning and reinforcement learning (RL), and adapts an off-the-shelf reward learning algorithm from RL for joint data manipulation learning and model training. Different parameterization of the "data reward" function instantiates different manipulation schemes. We showcase data augmentation that learns a text transformation network, and data weighting that dynamically adapts the data sample importance. Experiments show the resulting algorithms significantly improve the image and text classification performance in low data regime and class-imbalance problems.