Distribution-Balanced Loss for Multi-Label Classification in Long-Tailed Datasets

TL;DR

This paper introduces Distribution-Balanced Loss, a novel loss function designed to improve multi-label classification on long-tailed datasets by addressing label co-occurrence and negative label dominance, leading to better performance.

Contribution

The paper proposes a new loss function that re-balances weights considering label co-occurrence and includes negative tolerant regularization for multi-label long-tailed classification.

Findings

Significant performance improvements on Pascal VOC and COCO datasets.

Effective handling of label co-occurrence issues.

Mitigation of negative label over-suppression.

Abstract

We present a new loss function called Distribution-Balanced Loss for the multi-label recognition problems that exhibit long-tailed class distributions. Compared to conventional single-label classification problem, multi-label recognition problems are often more challenging due to two significant issues, namely the co-occurrence of labels and the dominance of negative labels (when treated as multiple binary classification problems). The Distribution-Balanced Loss tackles these issues through two key modifications to the standard binary cross-entropy loss: 1) a new way to re-balance the weights that takes into account the impact caused by label co-occurrence, and 2) a negative tolerant regularization to mitigate the over-suppression of negative labels. Experiments on both Pascal VOC and COCO show that the models trained with this new loss function achieve significant performance gains over existing methods. Code and models are available at: https://github.com/wutong16/DistributionBalancedLoss .