|
|
|
|
Left: building blocks for representation learning or classifier learning;
Right: overall abstraction of current works, where ‘→’ separates different stages of the method. Our work builds on GCD, and jointly trains a parametric classifier.
|
|
|
|
|
|
|
|
|
Code [GitHub] |
ICCV 2023 [Paper] |
Slides [Link] |
Poster [Link] |
|
|
|
|
|
Left: building blocks for representation learning or classifier learning;
Right: overall abstraction of current works, where ‘→’ separates different stages of the method. Our work builds on GCD, and jointly trains a parametric classifier.
|
|
Generalized Category Discovery (GCD) aims to discover novel categories in unlabelled datasets using knowledge learned from labelled samples. Previous studies argued that parametric classifiers are prone to overfitting to seen categories, and endorsed using a non-parametric classifier formed with semi-supervised k-means.
However, in this study, we investigate the failure of parametric clas- sifiers, verify the effectiveness of previous design choices when high-quality supervision is available, and identify unreliable pseudo-labels as a key problem. We demonstrate that two prediction biases exist: the classifier tends to predict seen classes more often, and produces an imbalanced distribution across seen and novel categories.
Based on these findings, we propose a simple yet effective parametric classification method that benefits from entropy regularisation, achieves state-of-the-art performance on multiple GCD benchmarks and shows strong robustness to unknown class numbers.
We hope the investigation and proposed simple framework can serve as a strong baseline to facilitate future studies in this field. Our code is available at: https://github.com/CVMI-Lab/SimGCD.
|
|
|
|
|
We observe strong biases in the prediction of the classifier, which are the main causes of the failure of parametric classifiers in GCD.
We then propose a simple yet effective parametric classification method that benefits from entropy regularisation, and achieves state-of-the-art performance on multiple GCD benchmarks.
|
|
|
Results on the generic image recognition datasets, the Semantic Shift Benchmark, and Herbarium 19.
|
|
|
Results with different numbers of categories.
Stronger entropy regularisation effectively enforces the model's robustness to unknown numbers of categories, but over-regularisation may limit the ability to recognise 'New' classes under ground-truth class numbers.
|
 |
Xin Wen, Bingchen Zhao, and Xiaojuan Qi Parametric Classification for Generalized Category Discovery: A Baseline Study In ICCV, 2023. |
|
@inproceedings{wen2023simgcd,
title={Parametric Classification for Generalized Category Discovery: A Baseline Study}, author={Wen, Xin and Zhao, Bingchen and Qi, Xiaojuan}, booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)}, year={2023}, pages={16590-16600} } |
Acknowledgements |