Rethinking Loss Reweighting for Imbalance Learning as an Inverse Problem: A Neural Collapse Point of View
Not on the map. Why?
Abstract (arXiv)
Loss reweighting is a widely used strategy for long-tailed classification, but existing reweighting strategies often rely on heuristics and rarely define a well-specified target. Inspired by Neural Collapse (NC), the ideal simplex Equiangular Tight Frame (ETF) terminal geometry suggests equal per-class average loss as a reasonable target for reweighting. Based on the ideal equal loss objective, we consider loss reweighting as an inverse problem and propose an inverse-view reweighting strategy that infers class weights dynamically to match this ideal objective. Empirically, NC metrics suggest our method can effectively reduce the loss imbalance coefficient and closer alignment with NC geometry while consistently outperforming strong long-tailed baselines on different datasets.
Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 30 of its 44 references that are in the map. First posted on arXiv 11 May 2026.
Builds on
30 of 44 references.
References · 30
- 2015Deep Residual Learning for Image Recognitioncited by 52,096
- 2019Class-Balanced Loss Based on Effective Number of Samplescited by 1,245
- 2019Decoupling Representation and Classifier for Long-Tailed Recognitioncited by 809
- 2020Prevalence of Neural Collapse during the terminal phase of deep learning trainingcited by 719
- 2020Long-tail learning via logit adjustmentcited by 618
- 2019BBN: Bilateral-Branch Network with Cumulative Learning for Long-Tailed Visual Recognitioncited by 403
- 2020Long-tailed Recognition by Routing Diverse Distribution-Aware Expertscited by 271
- 2021Improving Calibration for Long-Tailed Recognitioncited by 212
- 2020Rethinking Class-Balanced Methods for Long-Tailed Visual Recognition from a Domain Adaptation Perspectivecited by 143
- 2021Targeted Supervised Contrastive Learning for Long-Tailed Recognitioncited by 132
- 2021Contrastive Learning based Hybrid Networks for Long-Tailed Image Classificationcited by 130
- 2016Range Loss for Deep Face Recognition with Long-tailcited by 111
18 more
- 2022Long-Tailed Recognition via Weight Balancingcited by 107
- 2021The Majority Can Help The Minority: Context-rich Minority Oversampling for Long-tailed Classificationcited by 100
- 2022Inducing Neural Collapse in Imbalanced Learning: Do We Really Need a Learnable Classifier at the End of Deep Neural Network?cited by 100
- 2021Influence-Balanced Loss for Imbalanced Visual Classificationcited by 91
- 2022Are All Losses Created Equal: A Neural Collapse Perspectivecited by 78
- 2022Neural Collapse Inspired Attraction-Repulsion-Balanced Loss for Imbalanced Learningcited by 43
- 2023Global and Local Mixture Consistency Cumulative Learning for Long-tailed Visual Recognitionscited by 43
- 2019SoftAdapt: Techniques for Adaptive Loss Weighting of Neural Networks with Multi-Part Loss Functionscited by 40
- 2023Neural Collapse for Unconstrained Feature Model under Cross-entropy Loss with Imbalanced Datacited by 38
- 2024Gradient Reweighting: Towards Imbalanced Class-Incremental Learningcited by 37
- 2023Inducing Neural Collapse in Deep Long-tailed Learningcited by 28
- 2024Neural Collapse for Cross-entropy Class-Imbalanced Learning with Unconstrained ReLU Feature Modelcited by 19
- 2025Focal-SAM: Focal Sharpness-Aware Minimization for Long-Tailed Classificationcited by 14
- 2021Feature Generation for Long-tail Classificationcited by 9
- 2022Learning Sample Reweighting for Accuracy and Adversarial Robustnesscited by 6
- 2025You Are Your Own Best Teacher: Achieving Centralized-level Performance in Federated Learning under Heterogeneous and Long-tailed Datacited by 5
- 2025Space Alignment Matters: The Missing Piece for Inducing Neural Collapse in Long-Tailed Learningcited by 2
- 2025Long-Tailed Recognition via Information-Preservable Two-Stage Learningcited by 1
Built on by
Not cited yet.