Class Balancing Technique

Class balancing techniques aim to mitigate the negative impact of imbalanced datasets, where some classes have significantly fewer examples than others, on machine learning model performance. Current research focuses on improving existing methods like oversampling, undersampling, and loss weighting, exploring their effectiveness across various model architectures (including YOLO, SSD, Random Forests, and Generalized Linear Models) and investigating the interplay between class imbalance and other challenges like spurious correlations and noisy labels. These advancements are crucial for improving the fairness, robustness, and overall accuracy of machine learning models in diverse real-world applications, particularly in domains with inherently skewed data distributions such as object detection in aerial imagery or medical diagnosis.

Papers

July 19, 2024

The Group Robustness is in the Details: Revisiting Finetuning under Spurious Correlations
Tyler LaBonte, John C. Hill, Xinchen Zhang, Vidya Muthukumar, Abhishek Kumar
Spurious Correlation Covariance Matrix DetAIL Spectral Bias Group Robustness Worst Group Accuracy Class Balancing Technique

January 23, 2024

Enhancing Object Detection Performance for Small Objects through Synthetic Data Generation and Proportional Class-Balancing Technique: A Comparative Study in Industrial Scenarios
Jibinraj Antony, Vinit Hegiste, Ali Nazeri, Hooman Tavakoli, Snehal Walunj, Christiane Plociennik, Martin Ruskowski
Synthetic Data Object Detection Synthetic Data Generation Computer Vision Technique Tiny Object Detection Small Object Industrial Scenario Class Balancing Technique

December 17, 2023

Random Forest Variable Importance-based Selection Algorithm in Class Imbalance Problem
Yunbi Nam, Sunwoo Han
Random Forest Class Imbalance Variable Selection Selection Policy Variable Importance Class Balancing Technique

June 24, 2023

Evaluating the Utility of GAN Generated Synthetic Tabular Data for Class Balancing and Low Resource Settings
Nagarjuna Chereddy, Bharath Kumar Bolla
Training Data GAN Model Class Imbalance Task Utility GAN Based Synthetic Tabular Data Class Balancing Class Balancing Technique

March 22, 2023

Fairness Improves Learning from Noisily Labeled Long-Tailed Data
Jiaheng Wei, Zhaowei Zhu, Gang Niu, Tongliang Liu, Sijia Liu, Masashi Sugiyama, Yang Liu
LeArning Abstract Procedural Fairness Label Noise Long Tailed Data Fairness Regularization Performance Gap Class Balancing Technique

December 6, 2022

MUS-CDB: Mixed Uncertainty Sampling with Class Distribution Balancing for Active Annotation in Aerial Object Detection
Dong Liang, Jing-Wei Zhang, Ying-Peng Tang, Sheng-Jun Huang
Active Learning Aerial Object Detection Uncertainty Sampling Novel Active Learning Class Balancing Technique

January 8, 2022

Relieving Long-tailed Instance Segmentation via Pairwise Class Balance
Yin-Yin He, Peizhen Zhang, Xiu-Shen Wei, Xiangyu Zhang, Jian Sun
Long Tailed Pairwise Relationship Class Balancing Technique Fine Grained Bias

Class Balancing Technique

Papers

The Group Robustness is in the Details: Revisiting Finetuning under Spurious Correlations

Enhancing Object Detection Performance for Small Objects through Synthetic Data Generation and Proportional Class-Balancing Technique: A Comparative Study in Industrial Scenarios

Random Forest Variable Importance-based Selection Algorithm in Class Imbalance Problem

Evaluating the Utility of GAN Generated Synthetic Tabular Data for Class Balancing and Low Resource Settings

Fairness Improves Learning from Noisily Labeled Long-Tailed Data

MUS-CDB: Mixed Uncertainty Sampling with Class Distribution Balancing for Active Annotation in Aerial Object Detection

Relieving Long-tailed Instance Segmentation via Pairwise Class Balance