MESA: Boost Ensemble Imbalanced Learning with MEta-SAmpler

TL;DR

MESA introduces an adaptive ensemble learning framework for imbalanced data that learns sampling strategies directly from data, improving robustness and applicability over traditional heuristic methods.

Contribution

It proposes a novel ensemble imbalanced learning framework that decouples model training from meta-learning, enabling efficient and transferable sampling strategies.

Findings

MESA outperforms existing methods on synthetic and real-world datasets.

It demonstrates robustness and transferability across various tasks.

The framework is generally applicable to most learning models.

Abstract

Imbalanced learning (IL), i.e., learning unbiased models from class-imbalanced data, is a challenging problem. Typical IL methods including resampling and reweighting were designed based on some heuristic assumptions. They often suffer from unstable performance, poor applicability, and high computational cost in complex tasks where their assumptions do not hold. In this paper, we introduce a novel ensemble IL framework named MESA. It adaptively resamples the training set in iterations to get multiple classifiers and forms a cascade ensemble model. MESA directly learns the sampling strategy from data to optimize the final metric beyond following random heuristics. Moreover, unlike prevailing meta-learning-based IL solutions, we decouple the model-training and meta-training in MESA by independently train the meta-sampler over task-agnostic meta-data. This makes MESA generally applicable to most of the existing learning models and the meta-sampler can be efficiently applied to new tasks. Extensive experiments on both synthetic and real-world tasks demonstrate the effectiveness, robustness, and transferability of MESA. Our code is available at https://github.com/ZhiningLiu1998/mesa.