Trustworthy Long-Tailed Classification

Page view(s)
49
Checked on Jun 01, 2024
Trustworthy Long-Tailed Classification
Title:
Trustworthy Long-Tailed Classification
Journal Title:
2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
Keywords:
Publication Date:
27 September 2022
Citation:
Li, B., Han, Z., Li, H., Fu, H., & Zhang, C. (2022). Trustworthy Long-Tailed Classification. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). https://doi.org/10.1109/cvpr52688.2022.00684
Abstract:
Classification on long-tailed distributed data is a challenging problem, which suffers from serious class-imbalance and accordingly unpromising performance es-pecially on tail classes. Recently, the ensembling based methods achieve the state-of-the-art performance and show great potential. However, there are two limitations for cur-rent methods. First, their predictions are not trustworthy for failure-sensitive applications. This is especially harmful for the tail classes where the wrong predictions is basically fre-quent. Second, they assign unified numbers of experts to all samples, which is redundant for easy samples with excessive computational cost. To address these issues, we propose a Trustworthy Long-tailed Classification (TLC) method to jointly conduct classification and uncertainty estimation to identify hard samples in a multi-expert framework. Our TLC obtains the evidence-based uncertainty (EvU) and ev-idence for each expert, and then combines these uncer-tainties and evidences under the Dempster-Shafer Evidence Theory (DST). Moreover, we propose a dynamic expert en-gagement to reduce the number of engaged experts for easy samples and achieve efficiency while maintaining promising performances. Finally, we conduct comprehensive ex-periments on the tasks of classification, tail detection, OOD detection and failure prediction. The experimental results show that the proposed TLC outperforms existing methods and is trustworthy with reliable uncertainty.
License type:
Publisher Copyright
Funding Info:
This research / project is supported by the A*STAR - AI3 HTPO Seed Fund
Grant Reference no. : C211118012

This work was partly supported by the National Natural Science Foundation of China (61976151, 61732011), the National Key Research and Development Program of China under Grant 2019YFB2101900.
Description:
© 2022 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works.
ISSN:
1063-6919
Files uploaded:

File Size Format Action
trustworthy-long-tailed-classification.pdf 491.72 KB PDF Request a copy