Model-Agnostic Meta Learning for Class Imbalance Adaptation

TL;DR

HAMR is a flexible meta-learning framework that adaptively addresses class imbalance and data difficulty in NLP, improving minority class performance across diverse datasets.

Contribution

Introduces HAMR, a novel bi-level optimization-based method that dynamically weights challenging samples and amplifies hard example focus, outperforming existing baselines.

Findings

HAMR significantly improves minority class performance.

HAMR outperforms strong baseline methods across six datasets.

Ablation studies confirm the effectiveness of HAMR's modules.

Abstract

Class imbalance is a widespread challenge in NLP tasks, significantly hindering robust performance across diverse domains and applications. We introduce Hardness-Aware Meta-Resample (HAMR), a unified framework that adaptively addresses both class imbalance and data difficulty. HAMR employs bi-level optimizations to dynamically estimate instance-level weights that prioritize genuinely challenging samples and minority classes, while a neighborhood-aware resampling mechanism amplifies training focus on hard examples and their semantically similar neighbors. We validate HAMR on six imbalanced datasets covering multiple tasks and spanning biomedical, disaster response, and sentiment domains. Experimental results show that HAMR achieves substantial improvements for minority classes and consistently outperforms strong baselines. Extensive ablation studies demonstrate that our proposed modules synergistically contribute to performance gains and highlight HAMR as a flexible and generalizable approach for class imbalance adaptation. Code is available at https://github.com/trust-nlp/ImbalanceLearning.