Beyond re-balancing: distributionally robust augmentation against class-conditional distribution shift in long-tailed recognition

Keliang Li, Hong Chang, Shiguang Shan, Xilin CHEN

Keywords: Long-tailed recognition, data augmentation, distributionally robust optimization, im-balance

TL;DR: Study the problem of unreliable class-conditional distribution estimation in long-tailed recognition and propose a data augmentation method to solve it

Abstract:

As a fundamental and practical problem, long-tailed recognition has drawn burning attention. In this paper, we investigate an essential but rarely noticed issue in long-tailed recognition, Class-Conditional Distribution (CCD) shift due to scarce instances, which exhibits a significant discrepancy between the empirical CCDs for training and test data, especially for tail classes. We observe empirical evidence that the shift is a key factor that limits the performance of existing long-tailed learning methods, and provide novel understanding of these methods in the
course of our analysis. Motivated by this, we propose an adaptive data augmentation method, Distributionally Robust Augmentation (DRA), to learn models more robust to CCD shift. The convergence and generalization of DRA are theoretically guaranteed. Experimental results verify that DRA outperforms related data augmentation methods without extra training cost and significantly improves the performance of some existing long-tailed recognition methods.

Anonymous Url: I certify that there is no URL (e.g., github page) that could be used to find authors’ identity.

No Acknowledgement Section: I certify that there is no acknowledgement section in this submission for double blind review.

Code Of Ethics: I acknowledge that I and all co-authors of this work have read and commit to adhering to the ICLR Code of Ethics

Submission Guidelines: Yes

Please Choose The Closest Area That Your Submission Falls Into: Deep Learning and representational learning