Distribution-Balanced Loss for Multi-Label Classification in Long-Tailed Datasets
arXiv:2007.09654
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 .
arXiv commentTo appear in ECCV 2020 as a spotlight presentation. Code and models are available at: https://github.com/wutong16/DistributionBalancedLoss
References in corpus (10)
- 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
- Learning to Reweight Examples for Robust Deep Learning
- The Devil is in the Tails: Fine-grained Classification in the Wild
- What is the Effect of Importance Weighting in Deep Learning?
- Striking the Right Balance with Uncertainty
- Learning a Deep ConvNet for Multi-label Classification with Partial Labels
- MovieNet: A Holistic Dataset for Movie Understanding
- Open Compound Domain Adaptation