Posterior Re-calibration for Imbalanced Datasets

by   Junjiao Tian, et al.

Neural Networks can perform poorly when the training label distribution is heavily imbalanced, as well as when the testing data differs from the training distribution. In order to deal with shift in the testing label distribution, which imbalance causes, we motivate the problem from the perspective of an optimal Bayes classifier and derive a post-training prior rebalancing technique that can be solved through a KL-divergence based optimization. This method allows a flexible post-training hyper-parameter to be efficiently tuned on a validation set and effectively modify the classifier margin to deal with this imbalance. We further combine this method with existing likelihood shift methods, re-interpreting them from the same Bayesian perspective, and demonstrating that our method can deal with both problems in a unified way. The resulting algorithm can be conveniently used on probabilistic classification problems agnostic to underlying architectures. Our results on six different datasets and five different architectures show state of art accuracy, including on large-scale imbalanced datasets such as iNaturalist for classification and Synthia for semantic segmentation. Please see for implementation.


page 7

page 15


Phased Progressive Learning with Coupling-Regulation-Imbalance Loss for Imbalanced Classification

Deep neural networks generally perform poorly with datasets that suffer ...

Undoing the Damage of Label Shift for Cross-domain Semantic Segmentation

Existing works typically treat cross-domain semantic segmentation (CDSS)...

Mitigating Dataset Imbalance via Joint Generation and Classification

Supervised deep learning methods are enjoying enormous success in many p...

Improving Calibration for Long-Tailed Recognition

Deep neural networks may perform poorly when training datasets are heavi...

On Multi-Domain Long-Tailed Recognition, Generalization and Beyond

Real-world data often exhibit imbalanced label distributions. Existing s...

GCN-Based Linkage Prediction for Face Clustering on Imbalanced Datasets: An Empirical Study

In recent years, benefiting from the expressive power of Graph Convoluti...

The Devil is in the Margin: Margin-based Label Smoothing for Network Calibration

In spite of the dominant performances of deep neural networks, recent wo...