Optimal Subset Selection for Active Learning

Yifan Fu; Xingquan Zhu

Optimal Subset Selection for Active Learning

Yifan Fu, Xingquan Zhu

Research output: Book chapter/Published conference paper › Conference paper › peer-review

2 Citations (Scopus)

23 Downloads (Pure)

Abstract

Active learning traditionally relies on instance based utility measures to rank and select instances for labeling, which may result in labeling redundancy. To address this issue, we explore instance utility from twodimensions: individual uncertainty and instance disparity, using a correlation matrix. The active learning is transformed to a semi-definite programming problem toselect an optimal subset with maximum utility value. Experiments demonstrate the algorithm performance incomparison with baseline approaches.

Original language	English
Title of host publication	AAAI
Place of Publication	United States
Publisher	AAAI Press
Pages	1776-1777
Number of pages	2
Publication status	Published - 2011
Event	AAAI Conference on Artificial Intelligence - San Francisco, California, USA, New Zealand Duration: 07 Aug 2011 → 11 Aug 2011

Conference

Conference	AAAI Conference on Artificial Intelligence
Country/Territory	New Zealand
Period	07/08/11 → 11/08/11

Access to Document

1000002226postpub.pdfAccepted author manuscript, 488 KB

Cite this

@inproceedings{36f6d3859c264146a4cf0838fbff35c5,

title = "Optimal Subset Selection for Active Learning",

abstract = "Active learning traditionally relies on instance based utility measures to rank and select instances for labeling, which may result in labeling redundancy. To address this issue, we explore instance utility from twodimensions: individual uncertainty and instance disparity, using a correlation matrix. The active learning is transformed to a semi-definite programming problem toselect an optimal subset with maximum utility value. Experiments demonstrate the algorithm performance incomparison with baseline approaches.",

keywords = "Open access version available, Active learning, Classfication, Data mining",

author = "Yifan Fu and Xingquan Zhu",

note = "Imported on 16 May 2017 - DigiTool details were: publisher = United States: AAAI Press, 2011. Event dates (773o) = August 7-11, 2011; Parent title (773t) = AAAI Conference on Artificial Intelligence.; AAAI Conference on Artificial Intelligence ; Conference date: 07-08-2011 Through 11-08-2011",

year = "2011",

language = "English",

pages = "1776--1777",

booktitle = "AAAI",

publisher = "AAAI Press",

}

TY - GEN

T1 - Optimal Subset Selection for Active Learning

AU - Fu, Yifan

AU - Zhu, Xingquan

N1 - Imported on 16 May 2017 - DigiTool details were: publisher = United States: AAAI Press, 2011. Event dates (773o) = August 7-11, 2011; Parent title (773t) = AAAI Conference on Artificial Intelligence.

PY - 2011

Y1 - 2011

N2 - Active learning traditionally relies on instance based utility measures to rank and select instances for labeling, which may result in labeling redundancy. To address this issue, we explore instance utility from twodimensions: individual uncertainty and instance disparity, using a correlation matrix. The active learning is transformed to a semi-definite programming problem toselect an optimal subset with maximum utility value. Experiments demonstrate the algorithm performance incomparison with baseline approaches.

AB - Active learning traditionally relies on instance based utility measures to rank and select instances for labeling, which may result in labeling redundancy. To address this issue, we explore instance utility from twodimensions: individual uncertainty and instance disparity, using a correlation matrix. The active learning is transformed to a semi-definite programming problem toselect an optimal subset with maximum utility value. Experiments demonstrate the algorithm performance incomparison with baseline approaches.

KW - Open access version available

KW - Active learning

KW - Classfication

KW - Data mining

M3 - Conference paper

SP - 1776

EP - 1777

BT - AAAI

PB - AAAI Press

CY - United States

T2 - AAAI Conference on Artificial Intelligence

Y2 - 7 August 2011 through 11 August 2011

ER -

Optimal Subset Selection for Active Learning

Abstract

Conference

Access to Document

Fingerprint

Cite this