MetaBalance: High-Performance Neural Networks for Class-Imbalanced Data

2021/06/17 by Arpit Bansal, Bansal, Arpit, Micah Goldblum +9
Computer Science · Engineering · #Artificial Intelligence (cs.AI) #Domain Adaptation and Few-Shot Learning #FOS: Computer and information sciences #Imbalanced Data Classification Techniques #Vehicle License Plate Recognition

paper · pdf · doi:10.48550/arxiv.2106.09643

openalex publication_date 2021/06/17 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Class-imbalanced data, in which some classes contain far more samples than others, is ubiquitous in real-world applications. Standard techniques for handling class-imbalance usually work by training on a re-weighted loss or on re-balanced data. Unfortunately, training overparameterized neural networks on such objectives causes rapid memorization of minority class data. To avoid this trap, we harness meta-learning, which uses both an ''outer-loop'' and an ''inner-loop'' loss, each of which may be balanced using different strategies. We evaluate our method, MetaBalance, on image classification, credit-card fraud detection, loan default prediction, and facial recognition tasks with severely imbalanced data, and we find that MetaBalance outperforms a wide array of popular re-sampling strategies.

Citations

• Asymptotic Properties of Nearest Neighbor Rules Using Edited Data
• Learning Multiple Layers of Features from Tiny Images

Related

• MetaInfoNet: Learning Task-Guided Information for Sample Reweighting
• Imbalanced Classification via Explicit Gradient Learning From Augmented Data
• Introducing DeepBalance: Random Deep Belief Network Ensembles to Address Class Imbalance
• How Sensitive are Meta-Learners to Dataset Imbalance?
• Simplifying Neural Network Training Under Class Imbalance