Class-Balanced Loss Based on Effective Number of Samples

Abstract

With the rapid increase of large-scale, real-world datasets, it becomes critical to address the problem of long-tailed data distribution (i.e., a few classes account for most of the data, while most classes are under-represented). Existing solutions typically adopt class re-balancing strategies such as re-sampling and re-weighting based on the number of observations for each class. In this work, we argue that as the number of samples increases, the additional benefit of a newly added data point will diminish. We introduce a novel theoretical framework to measure data overlap by associating with each sample a small neighboring region rather than a single point. The effective number of samples is defined as the volume of samples and can be calculated by a simple formula $(1-{\beta }^n)/(1-\beta )$, where $n$ is the number of samples and $\beta \in [0,1)$ is a hyperparameter. We design a re-weighting scheme that uses the effective number of samples for each class to re-balance the loss, thereby yielding a class-balanced loss. Comprehensive experiments are conducted on artificially induced long-tailed CIFAR datasets and large-scale datasets including ImageNet and iNaturalist. Our results show that when trained with the proposed class-balanced loss, the network is able to achieve significant performance gains on long-tailed datasets.

arXiv commentCode is available at: https://github.com/richardaecn/class-balanced-loss

References in corpus (8)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• SMOTE: Synthetic Minority Over-sampling Technique
• Distributed Representations of Words and Phrases and their Compositionality
• A systematic study of the class imbalance problem in convolutional neural networks
• Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
• Understanding Black-box Predictions via Influence Functions
• The Devil is in the Tails: Fine-grained Classification in the Wild
• Deep Active Learning over the Long Tail

Cited by in corpus (38)

• Meta-Weight-Net: Learning an Explicit Mapping For Sample Weighting
• Global canopy height regression and uncertainty estimation from GEDI LIDAR waveforms with deep ensembles
• Blind Backdoors in Deep Learning Models
• BBN: Bilateral-Branch Network with Cumulative Learning for Long-Tailed Visual Recognition
• Data Augmentation for Object Detection via Progressive and Selective Instance-Switching
• Long-tail learning via logit adjustment
• CReST: A Class-Rebalancing Self-Training Framework for Imbalanced Semi-Supervised Learning
• Rethinking Class-Balanced Methods for Long-Tailed Visual Recognition from a Domain Adaptation Perspective
• The Herbarium Challenge 2019 Dataset
• Robustness May Be at Odds with Fairness: An Empirical Study on Class-wise Accuracy
• The iWildCam 2021 Competition Dataset
• MetaSAug: Meta Semantic Augmentation for Long-Tailed Visual Recognition
• The Devil is the Classifier: Investigating Long Tail Relation Classification with Decoupling Analysis
• Jointly De-biasing Face Recognition and Demographic Attribute Estimation
• Elastic-InfoGAN: Unsupervised Disentangled Representation Learning in Class-Imbalanced Data
• Semi-Supervised Semantic Segmentation via Adaptive Equalization Learning
• CondenseNeXt: An Ultra-Efficient Deep Neural Network for Embedded Systems
• End-2-End COVID-19 Detection from Breath & Cough Audio
• Coping with Label Shift via Distributionally Robust Optimisation
• Artificial Neural Networks and Guided Gene Expression Programming to Predict Wall Pressure Spectra Beneath Turbulent Boundary Layers
• Addressing Class Imbalance in Federated Learning
• Teacher's pet: understanding and mitigating biases in distillation
• Large-Scale Object Detection in the Wild from Imbalanced Multi-Labels
• SelectNet: Learning to Sample from the Wild for Imbalanced Data Training
• Disentangling Sampling and Labeling Bias for Learning in Large-Output Spaces
• Consistent Batch Normalization for Weighted Loss in Imbalanced-Data Environment
• DotFAN: A Domain-transferred Face Augmentation Network for Pose and Illumination Invariant Face Recognition
• Energy Aligning for Biased Models
• Improvement of Batch Normalization in Imbalanced Data
• Learning with Multiclass AUC: Theory and Algorithms
• Procrustean Training for Imbalanced Deep Learning
• Metadata Shaping: Natural Language Annotations for the Tail
• White Paper Assistance: A Step Forward Beyond the Shortcut Learning
• A Linkage-based Doubly Imbalanced Graph Learning Framework for Face Clustering
• Class Balancing GAN with a Classifier in the Loop
• Training Over-parameterized Models with Non-decomposable Objectives
• On-target Adaptation
• Collaging Class-specific GANs for Semantic Image Synthesis