Beyond Prototypes: Semantic Anchor Regularization for Better Representation Learning

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Ge, Yanqi, Nie, Qiang, Huang, Ye, Liu, Yong, Wang, Chengjie, Zheng, Feng, Li, Wen, Duan, Lixin
Format: Preprint
Publié: 2023
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866909092331651072
author Ge, Yanqi
Nie, Qiang
Huang, Ye
Liu, Yong
Wang, Chengjie
Zheng, Feng
Li, Wen
Duan, Lixin
author_facet Ge, Yanqi
Nie, Qiang
Huang, Ye
Liu, Yong
Wang, Chengjie
Zheng, Feng
Li, Wen
Duan, Lixin
contents One of the ultimate goals of representation learning is to achieve compactness within a class and well-separability between classes. Many outstanding metric-based and prototype-based methods following the Expectation-Maximization paradigm, have been proposed for this objective. However, they inevitably introduce biases into the learning process, particularly with long-tail distributed training data. In this paper, we reveal that the class prototype is not necessarily to be derived from training features and propose a novel perspective to use pre-defined class anchors serving as feature centroid to unidirectionally guide feature learning. However, the pre-defined anchors may have a large semantic distance from the pixel features, which prevents them from being directly applied. To address this issue and generate feature centroid independent from feature learning, a simple yet effective Semantic Anchor Regularization (SAR) is proposed. SAR ensures the interclass separability of semantic anchors in the semantic space by employing a classifier-aware auxiliary cross-entropy loss during training via disentanglement learning. By pulling the learned features to these semantic anchors, several advantages can be attained: 1) the intra-class compactness and naturally inter-class separability, 2) induced bias or errors from feature learning can be avoided, and 3) robustness to the long-tailed problem. The proposed SAR can be used in a plug-and-play manner in the existing models. Extensive experiments demonstrate that the SAR performs better than previous sophisticated prototype-based methods. The implementation is available at https://github.com/geyanqi/SAR.
format Preprint
id arxiv_https___arxiv_org_abs_2312_11872
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Beyond Prototypes: Semantic Anchor Regularization for Better Representation Learning
Ge, Yanqi
Nie, Qiang
Huang, Ye
Liu, Yong
Wang, Chengjie
Zheng, Feng
Li, Wen
Duan, Lixin
Computer Vision and Pattern Recognition
One of the ultimate goals of representation learning is to achieve compactness within a class and well-separability between classes. Many outstanding metric-based and prototype-based methods following the Expectation-Maximization paradigm, have been proposed for this objective. However, they inevitably introduce biases into the learning process, particularly with long-tail distributed training data. In this paper, we reveal that the class prototype is not necessarily to be derived from training features and propose a novel perspective to use pre-defined class anchors serving as feature centroid to unidirectionally guide feature learning. However, the pre-defined anchors may have a large semantic distance from the pixel features, which prevents them from being directly applied. To address this issue and generate feature centroid independent from feature learning, a simple yet effective Semantic Anchor Regularization (SAR) is proposed. SAR ensures the interclass separability of semantic anchors in the semantic space by employing a classifier-aware auxiliary cross-entropy loss during training via disentanglement learning. By pulling the learned features to these semantic anchors, several advantages can be attained: 1) the intra-class compactness and naturally inter-class separability, 2) induced bias or errors from feature learning can be avoided, and 3) robustness to the long-tailed problem. The proposed SAR can be used in a plug-and-play manner in the existing models. Extensive experiments demonstrate that the SAR performs better than previous sophisticated prototype-based methods. The implementation is available at https://github.com/geyanqi/SAR.
title Beyond Prototypes: Semantic Anchor Regularization for Better Representation Learning
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2312.11872