SIA OpenIR  > 机器人学研究室
What and How: Generalized Lifelong Spectral Clustering via Dual Memory
Sun G(孙干)1,2; Cong Y(丛杨)1,2; Dong JH(董家华)1,2,4; Liu YY(刘宇阳)1,2,4; Ding ZM(丁正明)3; Yu HB(于海斌)1,2
Department机器人学研究室
Source PublicationIEEE Transactions on Pattern Analysis and Machine Intelligence
ISSN0162-8828
2021
Pages1-13
Indexed ByEI
EI Accession number20210809936128
Contribution Rank1
Funding OrganizationNational Key Research and Development Program of China (2019YFB1310300) ; National Nature Science Foundation of China under Grant (62003336, 61821005, 61722311) ; National Postdoctoral Innovative Talents Support Program (BX20200353) ; Nature Foundation of Liaoning Province of China under Grant (2020-KF-11-01)
KeywordLifelong Machine Learning Spectral Clustering Deep Transfer Learning Neural Networks
Abstract

Spectral clustering has become one of the most effective clustering algorithms. We in this work explore the problem of spectral clustering in a lifelong learning framework termed as Generalized Lifelong Spectral Clustering (GL$^2$SC). Different from most current studies, which concentrate on a fixed spectral clustering task set and cannot efficiently incorporate a new clustering task, the goal of our work is to establish a generalized model for new spectral clustering task by What and How to lifelong learn from past tasks. For what to lifelong learn, our GL$^2$SC framework contains a dual memory mechanism with a deep orthogonal factorization manner: an orthogonal basis memory stores hidden and hierarchical clustering centers among learned tasks, and a feature embedding memory captures deep manifold representation common across multiple related tasks. When a new clustering task arrives, the intuition here for how to lifelong learn is that GL$^2$SC can transfer intrinsic knowledge from dual memory mechanism to obtain task-specific encoding matrix. Then the encoding matrix can redefine the dual memory over time to provide maximal benefits when learning future tasks. To the end, empirical comparisons on several benchmark datasets show the effectiveness of our GL$^2$SC, in comparison with several state-of-the-art spectral clustering models.

Language英语
Document Type期刊论文
Identifierhttp://ir.sia.cn/handle/173321/28328
Collection机器人学研究室
Corresponding AuthorCong Y(丛杨)
Affiliation1.State Key Laboratory of Robotics, Shenyang Institute of Automation, Chinese Academy of Sciences, shenyang, China
2.Institutes for Robotics and Intelligent Manufacturing, Chinese Academy of Sciences, Shenyang, China
3.Department of Computer Science, Tulane University, 5783 New Orleans, Louisiana, United States, 70118-5665
4.University of Chinese Academy of Sciences, Beijing, 100049, China
Recommended Citation
GB/T 7714
Sun G,Cong Y,Dong JH,et al. What and How: Generalized Lifelong Spectral Clustering via Dual Memory[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence,2021:1-13.
APA Sun G,Cong Y,Dong JH,Liu YY,Ding ZM,&Yu HB.(2021).What and How: Generalized Lifelong Spectral Clustering via Dual Memory.IEEE Transactions on Pattern Analysis and Machine Intelligence,1-13.
MLA Sun G,et al."What and How: Generalized Lifelong Spectral Clustering via Dual Memory".IEEE Transactions on Pattern Analysis and Machine Intelligence (2021):1-13.
Files in This Item:
File Name/Size DocType Version Access License
What and How_ Genera(14262KB)期刊论文作者接受稿开放获取CC BY-NC-SAView Application Full Text
Related Services
Recommend this item
Bookmark
Usage statistics
Export to Endnote
Google Scholar
Similar articles in Google Scholar
[Sun G(孙干)]'s Articles
[Cong Y(丛杨)]'s Articles
[Dong JH(董家华)]'s Articles
Baidu academic
Similar articles in Baidu academic
[Sun G(孙干)]'s Articles
[Cong Y(丛杨)]'s Articles
[Dong JH(董家华)]'s Articles
Bing Scholar
Similar articles in Bing Scholar
[Sun G(孙干)]'s Articles
[Cong Y(丛杨)]'s Articles
[Dong JH(董家华)]'s Articles
Terms of Use
No data!
Social Bookmark/Share
File name: What and How_ Generalized Lifelong Spectral Clustering via Dual Memory.pdf
Format: Adobe PDF
All comments (0)
No comment.
 

Items in the repository are protected by copyright, with all rights reserved, unless otherwise indicated.