Rethinking Loss Reweighting for Imbalance Learning as an Inverse Problem: A Neural Collapse Point of View

About

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.

Jinping Wang, Zixin Tong, Zhiwu Xie, Zhiqiang Gao · 2026

Related benchmarks

Task	Dataset	Result
Long-Tailed Image Classification	ImageNet-LT (test)	Top-1 Acc (Overall)57.6	246
Image Classification	CIFAR-10-LT (IF 50)	Top-1 Accuracy91.1	88
Image Classification	CIFAR-100 LT (IF=50)	Top-1 Acc63.9	82
Image Classification	CIFAR-10-LT IF 100	Top-1 Accuracy89.1	78
Image Classification	CIFAR-100 LT (IF=100)	Top-1 Acc58.6	55
Long-tailed recognition	iNaturalist 2018	Top-1 Accuracy74.9	31
Image Classification	CIFAR-10 LT (IF=200)	Top-1 Acc84	28
Image Classification	CIFAR-100-LT (IF=200)	Top-1 Acc52.4	27