M2m: Imbalanced Classification via Major-to-minor Translation

2020/04/01 by Jaehyung Kim, Jongheon Jeong, Kim, Jaehyung +3 · 18 citations
Computer Science · Mathematics · Medicine · #Artificial intelligence #Artificial neural network #COVID-19 diagnosis using AI #Class (philosophy) #Classifier (UML) #Computer science #Deep neural networks #Digital Imaging for Blood Diseases #Generalization #Imbalanced Data Classification Techniques #Machine learning #Mathematics #Pattern recognition (psychology) #Simple (philosophy) #Variety (cybernetics) #Weighting #cs.CV #cs.LG #stat.ML

paper · pdf · doi:10.48550/arxiv.2004.00431

published in arXiv (Cornell University) (Cornell University) · 12 pages; CVPR 2020

openalex publication_date 2020/04/01 · arxiv created 2020/12/20 · arxiv updated 2020/12/22 · openalex created_date 2025/10/10 · openalex updated_date 2026/08/08

Abstract

In most real-world scenarios, labeled training datasets are highly class-imbalanced, where deep neural networks suffer from generalizing to a balanced testing criterion. In this paper, we explore a novel yet simple way to alleviate this issue by augmenting less-frequent classes via translating samples (e.g., images) from more-frequent classes. This simple approach enables a classifier to learn more generalizable features of minority classes, by transferring and leveraging the diversity of the majority information. Our experimental results on a variety of class-imbalanced datasets show that the proposed method improves the generalization on minority classes significantly compared to other existing re-sampling or re-weighting methods. The performance of our method even surpasses those of previous state-of-the-art methods for the imbalanced classification.

Citations

• Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour

Cited by

• Balanced Learning for Domain Adaptive Semantic Segmentation
• Rectifying Soft-Label Entangled Bias in Long-Tailed Dataset Distillation
• LT-Soups: Bridging Head and Tail Classes via Subsampled Model Soups
• BlurGuard: A Simple Approach for Robustifying Image Protection Against AI-Powered Editing
• Distributional Robustness Loss for Long-tail Learning
• Learning Majority-to-Minority Transformations with MMD and Triplet Loss for Imbalanced Classification
• Feature-Space Oversampling for Addressing Class Imbalance in SAR Ship Classification
• Generative Active Learning for Long-tail Trajectory Prediction via Controllable Diffusion Model
• Long-Tailed Learning for Generalized Category Discovery
• Mixture Experts with Test-Time Self-Supervised Aggregation for Tabular Imbalanced Regression
• Breaking the Horizontal Prior: From Long-Tailed Orientation Bias to Roll-Robust Monocular Depth Estimation
• Focal-SAM: Focal Sharpness-Aware Minimization for Long-Tailed Classification

Related

• ReMix: Calibrated Resampling for Class Imbalance in Deep learning
• Imbalanced Classification via Explicit Gradient Learning From Augmented Data
• Effective Class-Imbalance learning based on SMOTE and Convolutional Neural Networks
• Imbalanced Classification via a Tabular Translation GAN
• Review of Methods for Handling Class-Imbalanced in Classification Problems