Abstract
Real-world data usually present long-tailed distributions. Training on imbalanced data tends to render neural networks perform well on head classes while much worse on tail classes. The severe sparseness of training instances for the tail classes is the main challenge, which results in biased distribution estimation during training. Plenty of efforts have been devoted to ameliorating the challenge, including data resampling and synthesizing new training instances for tail classes. However, no prior research has exploited the transferable knowledge from head classes to tail classes for calibrating the distribution of tail classes. In this article, we suppose that tail classes can be enriched by similar head classes and propose a novel distribution calibration (DC) approach named as l abel- a ware DC (). transfers the statistics from relevant head classes to infer the distribution of tail classes. Sampling from calibrated distribution further facilitates rebalancing the classifier. Experiments on both image and text long-tailed datasets demonstrate that significantly outperforms existing methods. The visualization also shows that provides a more accurate distribution estimation.
Funding Information
  • National Natural Science Foundation of China (62002084)
  • Stable Support Plan for Colleges and Universities in Shenzhen (GXWD2020 1230155427003-20200730101839009)
  • Guangdong Provincial Key Laboratory of Novel Security Intelligence Technologies (2022B1212010005)
  • Major Key Project of Peng Cheng Laboratory (PCL2022A03, PCL2021A02, PCL2021A09)