Data to 14 Oct 2026

Long-Tailed Classification by Keeping the Good and Removing the Bad Momentum Causal Effect

arXiv 2009.12991 ↗First posted 28 Sep 2020cs.CV cs.LG stat.MLcited by 256

Not on the map. Why?

Abstract (arXiv)

As the class size grows, maintaining a balanced dataset across many classes is challenging because the data are long-tailed in nature; it is even impossible when the sample-of-interest co-exists with each other in one collectable unit, e.g., multiple visual instances in one image. Therefore, long-tailed classification is the key to deep learning at scale. However, existing methods are mainly based on re-weighting/re-sampling heuristics that lack a fundamental theory. In this paper, we establish a causal inference framework, which not only unravels the whys of previous methods, but also derives a new principled solution. Specifically, our theory shows that the SGD momentum is essentially a confounder in long-tailed classification. On one hand, it has a harmful causal effect that misleads the tail prediction biased towards the head. On the other hand, its induced mediation also benefits the representation learning and head prediction. Our framework elegantly disentangles the paradoxical effects of the momentum, by pursuing the direct causal effect caused by an input sample. In particular, we use causal intervention in training, and counterfactual reasoning in inference, to remove the "bad" while keep the "good". We achieve new state-of-the-arts on three long-tailed visual recognition benchmarks: Long-tailed CIFAR-10/-100, ImageNet-LT for image classification and LVIS for instance segmentation.

Key facts. Cited by 256 papers in the map (data through 14 Oct 2026). First posted on arXiv 28 Sep 2020.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 66; 2022: 68; 2023: 52; 2024: 33; 2025: 24; 2026: 13.

256 papers.

Sort:

By area:

2021Deep Long-Tailed Learning: A Surveycited by 417
2021Parametric Contrastive Learningcited by 217
2021BENDR: using transformers and a contrastive self-supervised learning task to learn from massive amounts of EEG datacited by 199
2021Distribution Alignment: A Unified Framework for Long-tail Visual Recognitioncited by 198
2021CReST: A Class-Rebalancing Self-Training Framework for Imbalanced Semi-Supervised Learningcited by 175
2021Delving into Deep Imbalanced Regressioncited by 158
2022Balanced Contrastive Learning for Long-Tailed Visual Recognitioncited by 144
2021Contrastive Learning based Hybrid Networks for Long-Tailed Image Classificationcited by 130
2021Self-supervised Learning is More Robust to Dataset Imbalancecited by 123
2021Distilling Causal Effect of Data in Class-Incremental Learningcited by 117
2021Causal Attention for Interpretable and Generalizable Graph Classificationcited by 108
2022Long-Tailed Recognition via Weight Balancingcited by 107

38 more

2021The Majority Can Help The Minority: Context-rich Minority Oversampling for Long-tailed Classificationcited by 100
2021ResLT: Residual Learning for Long-tailed Recognitioncited by 96
2021Counterfactual Zero-Shot and Open-Set Visual Recognitioncited by 95
2021Self-Supervised Aggregation of Diverse Experts for Test-Agnostic Long-Tailed Recognitioncited by 95
2021Interventional Video Grounding with Dual Contrastive Learningcited by 94
2023Interventional Bag Multi-Instance Learning On Whole-Slide Pathological Imagescited by 91
2022Debiased Learning from Naturally Imbalanced Pseudo-Labelscited by 85
2021Causal Attention for Unbiased Visual Recognitioncited by 80
2021FASA: Feature Augmentation and Sampling Adaptation for Long-Tailed Instance Segmentationcited by 76
2022A Survey on Long-Tailed Visual Recognitioncited by 75
2023Long-tailed Visual Recognition via Gaussian Clouded Logit Adjustmentcited by 75
2021Self Supervision to Distillation for Long-Tailed Visual Recognitioncited by 73
2021Distilling Virtual Examples for Long-tailed Recognitioncited by 72
2021Learning Representation for Clustering via Prototype Scattering and Positive Samplingcited by 58
2022Cross-Modal Causal Relational Reasoning for Event-Level Visual Question Answeringcited by 57
2022Stop Wasting My Time! Saving Days of ImageNet and BERT Training with Latest Weight Averagingcited by 54
2022RankSim: Ranking Similarity Regularization for Deep Imbalanced Regressioncited by 52
2021Trustworthy Long-Tailed Classificationcited by 51
2021Adversarial Robustness under Long-Tailed Distributioncited by 50
2021A Simple Long-Tailed Recognition Baseline via Vision-Language Modelcited by 49
2022Equalized Focal Loss for Dense Long-Tailed Object Detectioncited by 46
2022Generalized Parametric Contrastive Learningcited by 45
2024Probabilistic Contrastive Learning for Long-Tailed Visual Recognitioncited by 42
2021CoSSL: Co-Learning of Representation and Classifier for Imbalanced Semi-Supervised Learningcited by 41
2022TAM: Topology-Aware Margin Loss for Class-Imbalanced Node Classificationcited by 41
2024The Clever Hans Mirage: A Comprehensive Survey on Spurious Correlations in Machine Learningcited by 41
2022Towards Long-Tailed 3D Detectioncited by 40
2021On Model Calibration for Long-Tailed Object Detection and Instance Segmentationcited by 39
2021Towards Calibrated Model for Long-Tailed Visual Recognition from Prior Perspectivecited by 39
2023Backdoor Defense via Deconfounded Representation Learningcited by 38
2021CausalAdv: Adversarial Robustness through the Lens of Causalitycited by 37
2021Robust Long-Tailed Learning under Label Noisecited by 36
2022Measuring Representational Harms in Image Captioningcited by 36
2021VideoLT: Large-scale Long-tailed Video Recognitioncited by 35
2023Exploring Vision-Language Models for Imbalanced Learningcited by 35
2021Analyzing Overfitting under Class Imbalance in Neural Networks for Image Segmentationcited by 34
2022Invariant Feature Learning for Generalized Long-Tailed Classificationcited by 34
2022Learning Imbalanced Data with Vision Transformerscited by 34

50 most cited of 256 All 256 (JSON)