Multi-instance learning with discriminative bag mapping

Jia Wu, Shirui Pan, Xingquan Zhu, Chengqi Zhang, Xindong Wu

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Multi-instance learning (MIL) is a useful tool for tackling labeling ambiguity in learning because it allows a bag of instances to share one label. Bag mapping transforms a bag into a single instance in a new space via instance selection and has drawn significant attention recently. To date, most existing work is based on the original space, using all instances inside each bag for bag mapping, and the selected instances are not directly tied to an MIL objective. As a result, it is difficult to guarantee the distinguishing capacity of the selected instances in the new bag mapping space. In this paper, we propose a discriminative mapping approach for multi-instance learning (MILDM) that aims to identify the best instances to directly distinguish bags in the new mapping space. Accordingly, each instance bag can be mapped using the selected instances to a new feature space, and hence any generic learning algorithm, such as an instance-based learning algorithm, can be used to derive learning models for multi-instance classification. Experiments and comparisons on eight different types of real-world learning tasks (including 14 data sets) demonstrate that MILDM outperforms the state-of-The-Art bag mapping multi-instance learning approaches. Results also confirm that MILDM achieves balanced performance between runtime efficiency and classification effectiveness.

Original languageEnglish
Pages (from-to)1065-1080
Number of pages16
JournalIEEE Transactions on Knowledge and Data Engineering
Volume30
Issue number6
DOIs
Publication statusPublished - 1 Jun 2018
Externally publishedYes

Keywords

  • Bag mapping
  • Classification
  • Instance selection
  • Multi-instance learning

Cite this

Wu, Jia ; Pan, Shirui ; Zhu, Xingquan ; Zhang, Chengqi ; Wu, Xindong. / Multi-instance learning with discriminative bag mapping. In: IEEE Transactions on Knowledge and Data Engineering. 2018 ; Vol. 30, No. 6. pp. 1065-1080.
@article{f45bad2b190a467787b68ae9cd7d609b,
title = "Multi-instance learning with discriminative bag mapping",
abstract = "Multi-instance learning (MIL) is a useful tool for tackling labeling ambiguity in learning because it allows a bag of instances to share one label. Bag mapping transforms a bag into a single instance in a new space via instance selection and has drawn significant attention recently. To date, most existing work is based on the original space, using all instances inside each bag for bag mapping, and the selected instances are not directly tied to an MIL objective. As a result, it is difficult to guarantee the distinguishing capacity of the selected instances in the new bag mapping space. In this paper, we propose a discriminative mapping approach for multi-instance learning (MILDM) that aims to identify the best instances to directly distinguish bags in the new mapping space. Accordingly, each instance bag can be mapped using the selected instances to a new feature space, and hence any generic learning algorithm, such as an instance-based learning algorithm, can be used to derive learning models for multi-instance classification. Experiments and comparisons on eight different types of real-world learning tasks (including 14 data sets) demonstrate that MILDM outperforms the state-of-The-Art bag mapping multi-instance learning approaches. Results also confirm that MILDM achieves balanced performance between runtime efficiency and classification effectiveness.",
keywords = "Bag mapping, Classification, Instance selection, Multi-instance learning",
author = "Jia Wu and Shirui Pan and Xingquan Zhu and Chengqi Zhang and Xindong Wu",
year = "2018",
month = "6",
day = "1",
doi = "10.1109/TKDE.2017.2788430",
language = "English",
volume = "30",
pages = "1065--1080",
journal = "IEEE Transactions on Knowledge and Data Engineering",
issn = "1041-4347",
publisher = "IEEE, Institute of Electrical and Electronics Engineers",
number = "6",

}

Multi-instance learning with discriminative bag mapping. / Wu, Jia; Pan, Shirui; Zhu, Xingquan; Zhang, Chengqi; Wu, Xindong.

In: IEEE Transactions on Knowledge and Data Engineering, Vol. 30, No. 6, 01.06.2018, p. 1065-1080.

Research output: Contribution to journalArticleResearchpeer-review

TY - JOUR

T1 - Multi-instance learning with discriminative bag mapping

AU - Wu, Jia

AU - Pan, Shirui

AU - Zhu, Xingquan

AU - Zhang, Chengqi

AU - Wu, Xindong

PY - 2018/6/1

Y1 - 2018/6/1

N2 - Multi-instance learning (MIL) is a useful tool for tackling labeling ambiguity in learning because it allows a bag of instances to share one label. Bag mapping transforms a bag into a single instance in a new space via instance selection and has drawn significant attention recently. To date, most existing work is based on the original space, using all instances inside each bag for bag mapping, and the selected instances are not directly tied to an MIL objective. As a result, it is difficult to guarantee the distinguishing capacity of the selected instances in the new bag mapping space. In this paper, we propose a discriminative mapping approach for multi-instance learning (MILDM) that aims to identify the best instances to directly distinguish bags in the new mapping space. Accordingly, each instance bag can be mapped using the selected instances to a new feature space, and hence any generic learning algorithm, such as an instance-based learning algorithm, can be used to derive learning models for multi-instance classification. Experiments and comparisons on eight different types of real-world learning tasks (including 14 data sets) demonstrate that MILDM outperforms the state-of-The-Art bag mapping multi-instance learning approaches. Results also confirm that MILDM achieves balanced performance between runtime efficiency and classification effectiveness.

AB - Multi-instance learning (MIL) is a useful tool for tackling labeling ambiguity in learning because it allows a bag of instances to share one label. Bag mapping transforms a bag into a single instance in a new space via instance selection and has drawn significant attention recently. To date, most existing work is based on the original space, using all instances inside each bag for bag mapping, and the selected instances are not directly tied to an MIL objective. As a result, it is difficult to guarantee the distinguishing capacity of the selected instances in the new bag mapping space. In this paper, we propose a discriminative mapping approach for multi-instance learning (MILDM) that aims to identify the best instances to directly distinguish bags in the new mapping space. Accordingly, each instance bag can be mapped using the selected instances to a new feature space, and hence any generic learning algorithm, such as an instance-based learning algorithm, can be used to derive learning models for multi-instance classification. Experiments and comparisons on eight different types of real-world learning tasks (including 14 data sets) demonstrate that MILDM outperforms the state-of-The-Art bag mapping multi-instance learning approaches. Results also confirm that MILDM achieves balanced performance between runtime efficiency and classification effectiveness.

KW - Bag mapping

KW - Classification

KW - Instance selection

KW - Multi-instance learning

UR - http://www.scopus.com/inward/record.url?scp=85040053375&partnerID=8YFLogxK

U2 - 10.1109/TKDE.2017.2788430

DO - 10.1109/TKDE.2017.2788430

M3 - Article

VL - 30

SP - 1065

EP - 1080

JO - IEEE Transactions on Knowledge and Data Engineering

JF - IEEE Transactions on Knowledge and Data Engineering

SN - 1041-4347

IS - 6

ER -