Improving Imbalanced Text Classification with Dynamic Curriculum Learning

TL;DR

This paper introduces a self-paced dynamic curriculum learning method that adaptively reorders training data based on difficulty, improving imbalanced text classification performance by mimicking human learning progression.

Contribution

The paper proposes a novel SPDCL approach that evaluates sample difficulty using linguistic features and model capacity, enabling adaptive curriculum scheduling for imbalanced datasets.

Findings

SPDCL outperforms static curriculum methods on multiple datasets.

The approach effectively handles class imbalance in text classification.

Adaptive reordering improves model learning efficiency.

Abstract

Recent advances in pre-trained language models have improved the performance for text classification tasks. However, little attention is paid to the priority scheduling strategy on the samples during training. Humans acquire knowledge gradually from easy to complex concepts, and the difficulty of the same material can also vary significantly in different learning stages. Inspired by this insights, we proposed a novel self-paced dynamic curriculum learning (SPDCL) method for imbalanced text classification, which evaluates the sample difficulty by both linguistic character and model capacity. Meanwhile, rather than using static curriculum learning as in the existing research, our SPDCL can reorder and resample training data by difficulty criterion with an adaptive from easy to hard pace. The extensive experiments on several classification tasks show the effectiveness of SPDCL strategy, especially for the imbalanced dataset.