HPCR: Holistic Proxy-based Contrastive Replay for Online Continual Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lin, Huiwei, Feng, Shanshan, Zhang, Baoquan, Li, Xutao, Ye, Yunming
Format: Preprint
Published: 2023
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929656109727744
author Lin, Huiwei
Feng, Shanshan
Zhang, Baoquan
Li, Xutao
Ye, Yunming
author_facet Lin, Huiwei
Feng, Shanshan
Zhang, Baoquan
Li, Xutao
Ye, Yunming
contents Online continual learning, aimed at developing a neural network that continuously learns new data from a single pass over an online data stream, generally suffers from catastrophic forgetting. Existing replay-based methods alleviate forgetting by replaying partial old data in a proxy-based or contrastive-based replay manner, each with its own shortcomings. Our previous work proposes a novel replay-based method called proxy-based contrastive replay (PCR), which handles the shortcomings by achieving complementary advantages of both replay manners. In this work, we further conduct gradient and limitation analysis of PCR. The analysis results show that PCR still can be further improved in feature extraction, generalization, and anti-forgetting capabilities of the model. Hence, we develop a more advanced method named holistic proxy-based contrastive replay (HPCR). HPCR consists of three components, each tackling one of the limitations of PCR. The contrastive component conditionally incorporates anchor-to-sample pairs to PCR, improving the feature extraction ability. The second is a temperature component that decouples the temperature coefficient into two parts based on their gradient impacts and sets different values for them to enhance the generalization ability. The third is a distillation component that constrains the learning process with additional loss terms to improve the anti-forgetting ability. Experiments on four datasets consistently demonstrate the superiority of HPCR over various state-of-the-art methods.
format Preprint
id arxiv_https___arxiv_org_abs_2309_15038
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle HPCR: Holistic Proxy-based Contrastive Replay for Online Continual Learning
Lin, Huiwei
Feng, Shanshan
Zhang, Baoquan
Li, Xutao
Ye, Yunming
Machine Learning
Computer Vision and Pattern Recognition
Online continual learning, aimed at developing a neural network that continuously learns new data from a single pass over an online data stream, generally suffers from catastrophic forgetting. Existing replay-based methods alleviate forgetting by replaying partial old data in a proxy-based or contrastive-based replay manner, each with its own shortcomings. Our previous work proposes a novel replay-based method called proxy-based contrastive replay (PCR), which handles the shortcomings by achieving complementary advantages of both replay manners. In this work, we further conduct gradient and limitation analysis of PCR. The analysis results show that PCR still can be further improved in feature extraction, generalization, and anti-forgetting capabilities of the model. Hence, we develop a more advanced method named holistic proxy-based contrastive replay (HPCR). HPCR consists of three components, each tackling one of the limitations of PCR. The contrastive component conditionally incorporates anchor-to-sample pairs to PCR, improving the feature extraction ability. The second is a temperature component that decouples the temperature coefficient into two parts based on their gradient impacts and sets different values for them to enhance the generalization ability. The third is a distillation component that constrains the learning process with additional loss terms to improve the anti-forgetting ability. Experiments on four datasets consistently demonstrate the superiority of HPCR over various state-of-the-art methods.
title HPCR: Holistic Proxy-based Contrastive Replay for Online Continual Learning
topic Machine Learning
Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2309.15038