Language-Inspired Relation Transfer for Few-Shot Class-Incremental Learning.

Yifan Zhao, Jia Li, Zeyin Song, Yonghong Tian
{"title":"Language-Inspired Relation Transfer for Few-Shot Class-Incremental Learning.","authors":"Yifan Zhao, Jia Li, Zeyin Song, Yonghong Tian","doi":"10.1109/TPAMI.2024.3492328","DOIUrl":null,"url":null,"abstract":"<p><p>Depicting novel classes with language descriptions by observing few-shot samples is inherent in human-learning systems. This lifelong learning capability helps to distinguish new knowledge from old ones through the increase of open-world learning, namely Few-Shot Class-Incremental Learning (FSCIL). Existing works to solve this problem mainly rely on the careful tuning of visual encoders, which shows an evident trade-off between the base knowledge and incremental ones. Motivated by human learning systems, we propose a new Language-inspired Relation Transfer (LRT) paradigm to understand objects by joint visual clues and text depictions, composed of two major steps. We first transfer the pretrained text knowledge to the visual domains by proposing a graph relation transformation module and then fuse the visual and language embedding by a text-vision prototypical fusion module. Second, to mitigate the domain gap caused by visual finetuning, we propose context prompt learning for fast domain alignment and imagined contrastive learning to alleviate the insufficient text data during alignment. With collaborative learning of domain alignments and text-image transfer, our proposed LRT outperforms the state-of-the-art models by over 13% and 7% on the final session of miniImageNet and CIFAR-100 FSCIL benchmarks.</p>","PeriodicalId":94034,"journal":{"name":"IEEE transactions on pattern analysis and machine intelligence","volume":"PP ","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-11-06","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"IEEE transactions on pattern analysis and machine intelligence","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/TPAMI.2024.3492328","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0

Abstract

Depicting novel classes with language descriptions by observing few-shot samples is inherent in human-learning systems. This lifelong learning capability helps to distinguish new knowledge from old ones through the increase of open-world learning, namely Few-Shot Class-Incremental Learning (FSCIL). Existing works to solve this problem mainly rely on the careful tuning of visual encoders, which shows an evident trade-off between the base knowledge and incremental ones. Motivated by human learning systems, we propose a new Language-inspired Relation Transfer (LRT) paradigm to understand objects by joint visual clues and text depictions, composed of two major steps. We first transfer the pretrained text knowledge to the visual domains by proposing a graph relation transformation module and then fuse the visual and language embedding by a text-vision prototypical fusion module. Second, to mitigate the domain gap caused by visual finetuning, we propose context prompt learning for fast domain alignment and imagined contrastive learning to alleviate the insufficient text data during alignment. With collaborative learning of domain alignments and text-image transfer, our proposed LRT outperforms the state-of-the-art models by over 13% and 7% on the final session of miniImageNet and CIFAR-100 FSCIL benchmarks.

语言启发下的关系转移,适用于少儿类增益学习。
通过观察少量样本来描绘具有语言描述的新类别是人类学习系统的固有特性。这种终身学习能力有助于通过增加开放世界学习(即 "少镜头类增量学习"(Few-Shot Class-Incremental Learning,FSCIL))来区分新旧知识。现有的解决这一问题的方法主要依赖于对视觉编码器的精心调整,这在基础知识和增量知识之间显示出明显的权衡。受人类学习系统的启发,我们提出了一种新的语言启发关系转移(LRT)范式,通过视觉线索和文本描述来理解物体,主要包括两个步骤。首先,我们通过提出图关系转换模块,将预先训练的文本知识转移到视觉领域,然后通过文本-视觉原型融合模块将视觉和语言嵌入融合在一起。其次,为了缓解视觉微调造成的领域差距,我们提出了上下文提示学习来实现快速领域对齐,并提出了想象对比学习来缓解对齐过程中文本数据不足的问题。通过领域配准和文本图像传输的协作学习,我们提出的 LRT 在 miniImageNet 和 CIFAR-100 FSCIL 基准的最终测试中分别以 13% 和 7% 的优势超越了最先进的模型。
本文章由计算机程序翻译,如有差异,请以英文原文为准。
求助全文
约1分钟内获得全文 求助全文
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