Active framework by sparsity exploitation for constructing a training set

Maozu Guo; Weining Wu; Yang Liu

doi:10.1007/978-3-319-95930-6_31

Active framework by sparsity exploitation for constructing a training set

Maozu Guo
, Weining Wu
, Yang Liu^*

^*Corresponding author for this work

Beijing University of Civil Engineering and Architecture
College of Computer Science and Technology, Harbin Engineering University
School of Computer Science and Technology, Harbin Institute of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

Abstract

This paper addresses the problem of actively constructing a training set for the linear model with sparse structure. This problem usually occurs in the scenario that no nonlinear mappings give similar performance for large-scale learning data, but it has to train a linear model quickly. In this paper, an active framework is proposed to reduce the time expense further in constructing the training set. The training examples are iteratively selected by matching partial components and their weights given by the classifier in pairs, in order to exploit model’s sparsity to precisely separate out more informative examples from others in a short time. The proposed framework is evaluated on a group of classification tasks, including the texts and images.

Original language	English
Title of host publication	Intelligent Computing - 14th International Conference, ICIC 2018, Proceedings
Editors	De-Shuang Huang, Vitoantonio Bevilacqua, Prashan Premaratne, Phalguni Gupta
Publisher	Springer Verlag
Pages	334-346
Number of pages	13
ISBN (Print)	9783319959290
DOIs	https://doi.org/10.1007/978-3-319-95930-6_31
State	Published - 2018
Externally published	Yes
Event	14th International Conference on Intelligent Computing, ICIC 2018 - Wuhan, China Duration: 15 Aug 2018 → 18 Aug 2018

Publication series

Name	Lecture Notes in Computer Science
Volume	10954 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	14th International Conference on Intelligent Computing, ICIC 2018
Country/Territory	China
City	Wuhan
Period	15/08/18 → 18/08/18

Keywords

Active framework
Classification
Sparse models

Access to Document

10.1007/978-3-319-95930-6_31

Cite this

Guo, M., Wu, W., & Liu, Y. (2018). Active framework by sparsity exploitation for constructing a training set. In D.-S. Huang, V. Bevilacqua, P. Premaratne, & P. Gupta (Eds.), Intelligent Computing - 14th International Conference, ICIC 2018, Proceedings (pp. 334-346). (Lecture Notes in Computer Science; Vol. 10954 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-95930-6_31

@inproceedings{3f7a457788bf4bcf99790b927eace65e,

title = "Active framework by sparsity exploitation for constructing a training set",

abstract = "This paper addresses the problem of actively constructing a training set for the linear model with sparse structure. This problem usually occurs in the scenario that no nonlinear mappings give similar performance for large-scale learning data, but it has to train a linear model quickly. In this paper, an active framework is proposed to reduce the time expense further in constructing the training set. The training examples are iteratively selected by matching partial components and their weights given by the classifier in pairs, in order to exploit model{\textquoteright}s sparsity to precisely separate out more informative examples from others in a short time. The proposed framework is evaluated on a group of classification tasks, including the texts and images.",

keywords = "Active framework, Classification, Sparse models",

author = "Maozu Guo and Weining Wu and Yang Liu",

note = "Publisher Copyright: {\textcopyright} 2018, Springer International Publishing AG, part of Springer Nature.; 14th International Conference on Intelligent Computing, ICIC 2018 ; Conference date: 15-08-2018 Through 18-08-2018",

year = "2018",

doi = "10.1007/978-3-319-95930-6\_31",

language = "英语",

isbn = "9783319959290",

series = "Lecture Notes in Computer Science",

publisher = "Springer Verlag",

pages = "334--346",

editor = "De-Shuang Huang and Vitoantonio Bevilacqua and Prashan Premaratne and Phalguni Gupta",

booktitle = "Intelligent Computing - 14th International Conference, ICIC 2018, Proceedings",

address = "德国",

}

Guo, M, Wu, W & Liu, Y 2018, Active framework by sparsity exploitation for constructing a training set. in D-S Huang, V Bevilacqua, P Premaratne & P Gupta (eds), Intelligent Computing - 14th International Conference, ICIC 2018, Proceedings. Lecture Notes in Computer Science, vol. 10954 LNCS, Springer Verlag, pp. 334-346, 14th International Conference on Intelligent Computing, ICIC 2018, Wuhan, China, 15/08/18. https://doi.org/10.1007/978-3-319-95930-6_31

Active framework by sparsity exploitation for constructing a training set. / Guo, Maozu; Wu, Weining; Liu, Yang.
Intelligent Computing - 14th International Conference, ICIC 2018, Proceedings. ed. / De-Shuang Huang; Vitoantonio Bevilacqua; Prashan Premaratne; Phalguni Gupta. Springer Verlag, 2018. p. 334-346 (Lecture Notes in Computer Science; Vol. 10954 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Active framework by sparsity exploitation for constructing a training set

AU - Guo, Maozu

AU - Wu, Weining

AU - Liu, Yang

PY - 2018

Y1 - 2018

N2 - This paper addresses the problem of actively constructing a training set for the linear model with sparse structure. This problem usually occurs in the scenario that no nonlinear mappings give similar performance for large-scale learning data, but it has to train a linear model quickly. In this paper, an active framework is proposed to reduce the time expense further in constructing the training set. The training examples are iteratively selected by matching partial components and their weights given by the classifier in pairs, in order to exploit model’s sparsity to precisely separate out more informative examples from others in a short time. The proposed framework is evaluated on a group of classification tasks, including the texts and images.

AB - This paper addresses the problem of actively constructing a training set for the linear model with sparse structure. This problem usually occurs in the scenario that no nonlinear mappings give similar performance for large-scale learning data, but it has to train a linear model quickly. In this paper, an active framework is proposed to reduce the time expense further in constructing the training set. The training examples are iteratively selected by matching partial components and their weights given by the classifier in pairs, in order to exploit model’s sparsity to precisely separate out more informative examples from others in a short time. The proposed framework is evaluated on a group of classification tasks, including the texts and images.

KW - Active framework

KW - Classification

KW - Sparse models

UR - https://www.scopus.com/pages/publications/85051872065

U2 - 10.1007/978-3-319-95930-6_31

DO - 10.1007/978-3-319-95930-6_31

M3 - 会议稿件

AN - SCOPUS:85051872065

SN - 9783319959290

T3 - Lecture Notes in Computer Science

SP - 334

EP - 346

BT - Intelligent Computing - 14th International Conference, ICIC 2018, Proceedings

A2 - Huang, De-Shuang

A2 - Bevilacqua, Vitoantonio

A2 - Premaratne, Prashan

A2 - Gupta, Phalguni

PB - Springer Verlag

T2 - 14th International Conference on Intelligent Computing, ICIC 2018

Y2 - 15 August 2018 through 18 August 2018

ER -

Active framework by sparsity exploitation for constructing a training set

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this