A transfer learning approach for network modeling

Shuai Huang, Jing Li, Kewei Chen, Teresa Wu, Jieping Ye, Xia Wu, Li Yao

Research output: Contribution to journalArticle

13 Citations (Scopus)

Abstract

Network models have been widely used in many subject areas to characterize the interactions between physical entities. A typical problem is to identify the network for multiple related tasks that share some similarities. In this case, a transfer learning approach that can leverage the knowledge gained during the modeling of one task to help better model another task is highly desirable. This article proposes a transfer learning approach that adopts a Bayesian hierarchical model framework to characterize the relatedness between tasks and additionally uses L 1-regularization to ensure robust learning of the networks with limited sample sizes. A method based on the Expectation- Maximization (EM) algorithm is further developed to learn the networks from data. Simulation studies are performed that demonstrate the superiority of the proposed transfer learning approach over single-task learning that learns the network of each task in isolation. The proposed approach is also applied to identify brain connectivity networks associated with Alzheimers Disease (AD) from functional magnetic resonance image data. The findings are consistent with the AD literature.

Original languageEnglish (US)
Pages (from-to)915-931
Number of pages17
JournalIIE Transactions (Institute of Industrial Engineers)
Volume44
Issue number11
DOIs
StatePublished - Nov 1 2012

Fingerprint

Magnetic resonance
Brain

Keywords

  • Bayesian hierarchical models
  • brain networks
  • fMRI
  • graphical models
  • Transfer learning

ASJC Scopus subject areas

  • Industrial and Manufacturing Engineering

Cite this

A transfer learning approach for network modeling. / Huang, Shuai; Li, Jing; Chen, Kewei; Wu, Teresa; Ye, Jieping; Wu, Xia; Yao, Li.

In: IIE Transactions (Institute of Industrial Engineers), Vol. 44, No. 11, 01.11.2012, p. 915-931.

Research output: Contribution to journalArticle

Huang, Shuai ; Li, Jing ; Chen, Kewei ; Wu, Teresa ; Ye, Jieping ; Wu, Xia ; Yao, Li. / A transfer learning approach for network modeling. In: IIE Transactions (Institute of Industrial Engineers). 2012 ; Vol. 44, No. 11. pp. 915-931.
@article{e93ae2a63aed43dbb12d1e3dc9bc6760,
title = "A transfer learning approach for network modeling",
abstract = "Network models have been widely used in many subject areas to characterize the interactions between physical entities. A typical problem is to identify the network for multiple related tasks that share some similarities. In this case, a transfer learning approach that can leverage the knowledge gained during the modeling of one task to help better model another task is highly desirable. This article proposes a transfer learning approach that adopts a Bayesian hierarchical model framework to characterize the relatedness between tasks and additionally uses L 1-regularization to ensure robust learning of the networks with limited sample sizes. A method based on the Expectation- Maximization (EM) algorithm is further developed to learn the networks from data. Simulation studies are performed that demonstrate the superiority of the proposed transfer learning approach over single-task learning that learns the network of each task in isolation. The proposed approach is also applied to identify brain connectivity networks associated with Alzheimers Disease (AD) from functional magnetic resonance image data. The findings are consistent with the AD literature.",
keywords = "Bayesian hierarchical models, brain networks, fMRI, graphical models, Transfer learning",
author = "Shuai Huang and Jing Li and Kewei Chen and Teresa Wu and Jieping Ye and Xia Wu and Li Yao",
year = "2012",
month = "11",
day = "1",
doi = "10.1080/0740817X.2011.649390",
language = "English (US)",
volume = "44",
pages = "915--931",
journal = "IISE Transactions",
issn = "2472-5854",
publisher = "Taylor and Francis Ltd.",
number = "11",

}

TY - JOUR

T1 - A transfer learning approach for network modeling

AU - Huang, Shuai

AU - Li, Jing

AU - Chen, Kewei

AU - Wu, Teresa

AU - Ye, Jieping

AU - Wu, Xia

AU - Yao, Li

PY - 2012/11/1

Y1 - 2012/11/1

N2 - Network models have been widely used in many subject areas to characterize the interactions between physical entities. A typical problem is to identify the network for multiple related tasks that share some similarities. In this case, a transfer learning approach that can leverage the knowledge gained during the modeling of one task to help better model another task is highly desirable. This article proposes a transfer learning approach that adopts a Bayesian hierarchical model framework to characterize the relatedness between tasks and additionally uses L 1-regularization to ensure robust learning of the networks with limited sample sizes. A method based on the Expectation- Maximization (EM) algorithm is further developed to learn the networks from data. Simulation studies are performed that demonstrate the superiority of the proposed transfer learning approach over single-task learning that learns the network of each task in isolation. The proposed approach is also applied to identify brain connectivity networks associated with Alzheimers Disease (AD) from functional magnetic resonance image data. The findings are consistent with the AD literature.

AB - Network models have been widely used in many subject areas to characterize the interactions between physical entities. A typical problem is to identify the network for multiple related tasks that share some similarities. In this case, a transfer learning approach that can leverage the knowledge gained during the modeling of one task to help better model another task is highly desirable. This article proposes a transfer learning approach that adopts a Bayesian hierarchical model framework to characterize the relatedness between tasks and additionally uses L 1-regularization to ensure robust learning of the networks with limited sample sizes. A method based on the Expectation- Maximization (EM) algorithm is further developed to learn the networks from data. Simulation studies are performed that demonstrate the superiority of the proposed transfer learning approach over single-task learning that learns the network of each task in isolation. The proposed approach is also applied to identify brain connectivity networks associated with Alzheimers Disease (AD) from functional magnetic resonance image data. The findings are consistent with the AD literature.

KW - Bayesian hierarchical models

KW - brain networks

KW - fMRI

KW - graphical models

KW - Transfer learning

UR - http://www.scopus.com/inward/record.url?scp=84865856155&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84865856155&partnerID=8YFLogxK

U2 - 10.1080/0740817X.2011.649390

DO - 10.1080/0740817X.2011.649390

M3 - Article

AN - SCOPUS:84865856155

VL - 44

SP - 915

EP - 931

JO - IISE Transactions

JF - IISE Transactions

SN - 2472-5854

IS - 11

ER -