利用重要性采样和原型-实例关系提炼进行对比式持续学习

ArXiv Pub Date : 2024-03-07 DOI:10.1609/aaai.v38i12.29259
Jiyong Li, Dilshod Azizov, Yang Li, Shangsong Liang
{"title":"利用重要性采样和原型-实例关系提炼进行对比式持续学习","authors":"Jiyong Li, Dilshod Azizov, Yang Li, Shangsong Liang","doi":"10.1609/aaai.v38i12.29259","DOIUrl":null,"url":null,"abstract":"Recently, because of the high-quality representations of contrastive learning methods, rehearsal-based contrastive continual learning has been proposed to explore how to continually learn transferable representation embeddings to avoid the catastrophic forgetting issue in traditional continual settings. Based on this framework, we propose Contrastive Continual Learning via Importance Sampling (CCLIS) to preserve knowledge by recovering previous data distributions with a new strategy for Replay Buffer Selection (RBS), which minimize estimated variance to save hard negative samples for representation learning with high quality. Furthermore, we present the Prototype-instance Relation Distillation (PRD) loss, a technique designed to maintain the relationship between prototypes and sample representations using a self-distillation process. Experiments on standard continual learning benchmarks reveal that our method notably outperforms existing baselines in terms of knowledge preservation and thereby effectively counteracts catastrophic forgetting in online contexts. The code is available at https://github.com/lijy373/CCLIS.","PeriodicalId":513202,"journal":{"name":"ArXiv","volume":"25 46","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-03-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Contrastive Continual Learning with Importance Sampling and Prototype-Instance Relation Distillation\",\"authors\":\"Jiyong Li, Dilshod Azizov, Yang Li, Shangsong Liang\",\"doi\":\"10.1609/aaai.v38i12.29259\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Recently, because of the high-quality representations of contrastive learning methods, rehearsal-based contrastive continual learning has been proposed to explore how to continually learn transferable representation embeddings to avoid the catastrophic forgetting issue in traditional continual settings. Based on this framework, we propose Contrastive Continual Learning via Importance Sampling (CCLIS) to preserve knowledge by recovering previous data distributions with a new strategy for Replay Buffer Selection (RBS), which minimize estimated variance to save hard negative samples for representation learning with high quality. Furthermore, we present the Prototype-instance Relation Distillation (PRD) loss, a technique designed to maintain the relationship between prototypes and sample representations using a self-distillation process. Experiments on standard continual learning benchmarks reveal that our method notably outperforms existing baselines in terms of knowledge preservation and thereby effectively counteracts catastrophic forgetting in online contexts. The code is available at https://github.com/lijy373/CCLIS.\",\"PeriodicalId\":513202,\"journal\":{\"name\":\"ArXiv\",\"volume\":\"25 46\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-03-07\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"ArXiv\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1609/aaai.v38i12.29259\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"ArXiv","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1609/aaai.v38i12.29259","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

摘要

最近,由于对比学习方法具有高质量的表征,人们提出了基于排演的对比持续学习,以探索如何持续学习可转移的表征嵌入,从而避免传统持续学习设置中的灾难性遗忘问题。在此框架基础上,我们提出了通过重要度采样进行对比式连续学习(CCLIS),通过一种新的重放缓冲区选择(RBS)策略恢复以前的数据分布,从而保存知识,这种策略能最大限度地减少估计方差,为高质量的表征学习保存硬负样本。此外,我们还提出了原型-实例关系蒸馏(PRD)损失,这是一种旨在利用自蒸馏过程保持原型与样本表示之间关系的技术。在标准的持续学习基准上进行的实验表明,我们的方法在知识保存方面明显优于现有的基准,从而有效地抵消了在线环境下的灾难性遗忘。代码见 https://github.com/lijy373/CCLIS。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Contrastive Continual Learning with Importance Sampling and Prototype-Instance Relation Distillation
Recently, because of the high-quality representations of contrastive learning methods, rehearsal-based contrastive continual learning has been proposed to explore how to continually learn transferable representation embeddings to avoid the catastrophic forgetting issue in traditional continual settings. Based on this framework, we propose Contrastive Continual Learning via Importance Sampling (CCLIS) to preserve knowledge by recovering previous data distributions with a new strategy for Replay Buffer Selection (RBS), which minimize estimated variance to save hard negative samples for representation learning with high quality. Furthermore, we present the Prototype-instance Relation Distillation (PRD) loss, a technique designed to maintain the relationship between prototypes and sample representations using a self-distillation process. Experiments on standard continual learning benchmarks reveal that our method notably outperforms existing baselines in terms of knowledge preservation and thereby effectively counteracts catastrophic forgetting in online contexts. The code is available at https://github.com/lijy373/CCLIS.
求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
自引率
0.00%
发文量
0
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信