Abstract

In many real world applications such as image annotation, gene function prediction, and insider threat detection, the data collected from heterogeneous sources often exhibit multiple types of heterogeneity, such as task heterogeneity, view heterogeneity, and label heterogeneity. To address this problem, we propose a Hierarchical Multi-Latent Space (HiMLS) learning framework to jointly model the triple types of heterogeneity. The basic idea is to learn a hierarchical multi-latent space by which we can simultaneously leverage the task relatedness, view consistency and the label correlations to improve the learning performance. We first propose a multi-latent space approach to model the complex heterogeneity, which is then used as a building block to stack up a multi-layer structure in order to learn the hierarchical multi-latent space. In such a way, we can gradually learn the more abstract concepts in the higher level. We present two instantiated models of the generalized framework using different divergence measures. The two-phase learning algorithms are used to train the multi-layer models. We drive the multiplicative update rules for pre-training and fine-tuning in each model, and prove the convergence and correctness of the update methods. The effectiveness of the proposed approach is verified on various data sets.

Original languageEnglish (US)
Article number7572203
Pages (from-to)3154-3168
Number of pages15
JournalIEEE Transactions on Knowledge and Data Engineering
Volume28
Issue number12
DOIs
StatePublished - Dec 1 2016

Fingerprint

Labels
Learning algorithms
Tuning
Genes

Keywords

  • Heterogeneous learning
  • matrix tri-factorization
  • multi-label learning
  • multi-task learning
  • multi-view learning

ASJC Scopus subject areas

  • Information Systems
  • Computer Science Applications
  • Computational Theory and Mathematics

Cite this

A Generalized Hierarchical Multi-Latent Space Model for Heterogeneous Learning. / Yang, Pei; Davulcu, Hasan; Zhu, Yada; He, Jingrui.

In: IEEE Transactions on Knowledge and Data Engineering, Vol. 28, No. 12, 7572203, 01.12.2016, p. 3154-3168.

Research output: Contribution to journalArticle

@article{97e0475514344d7486cd92335625d134,
title = "A Generalized Hierarchical Multi-Latent Space Model for Heterogeneous Learning",
abstract = "In many real world applications such as image annotation, gene function prediction, and insider threat detection, the data collected from heterogeneous sources often exhibit multiple types of heterogeneity, such as task heterogeneity, view heterogeneity, and label heterogeneity. To address this problem, we propose a Hierarchical Multi-Latent Space (HiMLS) learning framework to jointly model the triple types of heterogeneity. The basic idea is to learn a hierarchical multi-latent space by which we can simultaneously leverage the task relatedness, view consistency and the label correlations to improve the learning performance. We first propose a multi-latent space approach to model the complex heterogeneity, which is then used as a building block to stack up a multi-layer structure in order to learn the hierarchical multi-latent space. In such a way, we can gradually learn the more abstract concepts in the higher level. We present two instantiated models of the generalized framework using different divergence measures. The two-phase learning algorithms are used to train the multi-layer models. We drive the multiplicative update rules for pre-training and fine-tuning in each model, and prove the convergence and correctness of the update methods. The effectiveness of the proposed approach is verified on various data sets.",
keywords = "Heterogeneous learning, matrix tri-factorization, multi-label learning, multi-task learning, multi-view learning",
author = "Pei Yang and Hasan Davulcu and Yada Zhu and Jingrui He",
year = "2016",
month = "12",
day = "1",
doi = "10.1109/TKDE.2016.2611514",
language = "English (US)",
volume = "28",
pages = "3154--3168",
journal = "IEEE Transactions on Knowledge and Data Engineering",
issn = "1041-4347",
publisher = "IEEE Computer Society",
number = "12",

}

TY - JOUR

T1 - A Generalized Hierarchical Multi-Latent Space Model for Heterogeneous Learning

AU - Yang, Pei

AU - Davulcu, Hasan

AU - Zhu, Yada

AU - He, Jingrui

PY - 2016/12/1

Y1 - 2016/12/1

N2 - In many real world applications such as image annotation, gene function prediction, and insider threat detection, the data collected from heterogeneous sources often exhibit multiple types of heterogeneity, such as task heterogeneity, view heterogeneity, and label heterogeneity. To address this problem, we propose a Hierarchical Multi-Latent Space (HiMLS) learning framework to jointly model the triple types of heterogeneity. The basic idea is to learn a hierarchical multi-latent space by which we can simultaneously leverage the task relatedness, view consistency and the label correlations to improve the learning performance. We first propose a multi-latent space approach to model the complex heterogeneity, which is then used as a building block to stack up a multi-layer structure in order to learn the hierarchical multi-latent space. In such a way, we can gradually learn the more abstract concepts in the higher level. We present two instantiated models of the generalized framework using different divergence measures. The two-phase learning algorithms are used to train the multi-layer models. We drive the multiplicative update rules for pre-training and fine-tuning in each model, and prove the convergence and correctness of the update methods. The effectiveness of the proposed approach is verified on various data sets.

AB - In many real world applications such as image annotation, gene function prediction, and insider threat detection, the data collected from heterogeneous sources often exhibit multiple types of heterogeneity, such as task heterogeneity, view heterogeneity, and label heterogeneity. To address this problem, we propose a Hierarchical Multi-Latent Space (HiMLS) learning framework to jointly model the triple types of heterogeneity. The basic idea is to learn a hierarchical multi-latent space by which we can simultaneously leverage the task relatedness, view consistency and the label correlations to improve the learning performance. We first propose a multi-latent space approach to model the complex heterogeneity, which is then used as a building block to stack up a multi-layer structure in order to learn the hierarchical multi-latent space. In such a way, we can gradually learn the more abstract concepts in the higher level. We present two instantiated models of the generalized framework using different divergence measures. The two-phase learning algorithms are used to train the multi-layer models. We drive the multiplicative update rules for pre-training and fine-tuning in each model, and prove the convergence and correctness of the update methods. The effectiveness of the proposed approach is verified on various data sets.

KW - Heterogeneous learning

KW - matrix tri-factorization

KW - multi-label learning

KW - multi-task learning

KW - multi-view learning

UR - http://www.scopus.com/inward/record.url?scp=84996551351&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84996551351&partnerID=8YFLogxK

U2 - 10.1109/TKDE.2016.2611514

DO - 10.1109/TKDE.2016.2611514

M3 - Article

VL - 28

SP - 3154

EP - 3168

JO - IEEE Transactions on Knowledge and Data Engineering

JF - IEEE Transactions on Knowledge and Data Engineering

SN - 1041-4347

IS - 12

M1 - 7572203

ER -