Multi-source multi-target dictionary learning for prediction of cognitive decline

Jie Zhang, Qingyang Li, Richard John Caselli, Paul M. Thompson, Jieping Ye, Yalin Wang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Alzheimer’s Disease (AD) is the most common type of dementia. Identifying correct biomarkers may determine presymptomatic AD subjects and enable early intervention. Recently, Multitask sparse feature learning has been successfully applied to many computer vision and biomedical informatics researches. It aims to improve the generalization performance by exploiting the shared features among different tasks. However, most of the existing algorithms are formulated as a supervised learning scheme. Its drawback is with either insufficient feature numbers or missing label information. To address these challenges, we formulate an unsupervised framework for multi-task sparse feature learning based on a novel dictionary learning algorithm. To solve the unsupervised learning problem, we propose a two-stage Multi-Source Multi-Target Dictionary Learning (MMDL) algorithm. In stage 1, we propose a multi-source dictionary learning method to utilize the common and individual sparse features in different time slots. In stage 2, supported by a rigorous theoretical analysis, we develop a multi-task learning method to solve the missing label problem. Empirical studies on an N = 3970 longitudinal brain image data set, which involves 2 sources and 5 targets, demonstrate the improved prediction accuracy and speed efficiency of MMDL in comparison with other state-of-the-art algorithms.

Original languageEnglish (US)
Title of host publicationInformation Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings
PublisherSpringer Verlag
Pages184-197
Number of pages14
Volume10265 LNCS
ISBN (Print)9783319590493
DOIs
StatePublished - 2017
Event25th International Conference on Information Processing in Medical Imaging, IPMI 2017 - Boone, United States
Duration: Jun 25 2017Jun 30 2017

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume10265 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other25th International Conference on Information Processing in Medical Imaging, IPMI 2017
CountryUnited States
CityBoone
Period6/25/176/30/17

Fingerprint

Glossaries
Target
Alzheimer's Disease
Prediction
Learning algorithms
Labels
Learning Algorithm
Multi-task Learning
Dementia
Unsupervised learning
Unsupervised Learning
Supervised learning
Biomarkers
Supervised Learning
Computer Vision
Empirical Study
Computer vision
Brain
Theoretical Analysis
Learning

Keywords

  • Alzheimer’s disease
  • Dictionary learning
  • Multi-task

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Zhang, J., Li, Q., Caselli, R. J., Thompson, P. M., Ye, J., & Wang, Y. (2017). Multi-source multi-target dictionary learning for prediction of cognitive decline. In Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings (Vol. 10265 LNCS, pp. 184-197). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10265 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-319-59050-9_15

Multi-source multi-target dictionary learning for prediction of cognitive decline. / Zhang, Jie; Li, Qingyang; Caselli, Richard John; Thompson, Paul M.; Ye, Jieping; Wang, Yalin.

Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings. Vol. 10265 LNCS Springer Verlag, 2017. p. 184-197 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 10265 LNCS).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Zhang, J, Li, Q, Caselli, RJ, Thompson, PM, Ye, J & Wang, Y 2017, Multi-source multi-target dictionary learning for prediction of cognitive decline. in Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings. vol. 10265 LNCS, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 10265 LNCS, Springer Verlag, pp. 184-197, 25th International Conference on Information Processing in Medical Imaging, IPMI 2017, Boone, United States, 6/25/17. https://doi.org/10.1007/978-3-319-59050-9_15
Zhang J, Li Q, Caselli RJ, Thompson PM, Ye J, Wang Y. Multi-source multi-target dictionary learning for prediction of cognitive decline. In Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings. Vol. 10265 LNCS. Springer Verlag. 2017. p. 184-197. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-59050-9_15
Zhang, Jie ; Li, Qingyang ; Caselli, Richard John ; Thompson, Paul M. ; Ye, Jieping ; Wang, Yalin. / Multi-source multi-target dictionary learning for prediction of cognitive decline. Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings. Vol. 10265 LNCS Springer Verlag, 2017. pp. 184-197 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{e68ecb27d643450883abf1912decd90c,
title = "Multi-source multi-target dictionary learning for prediction of cognitive decline",
abstract = "Alzheimer’s Disease (AD) is the most common type of dementia. Identifying correct biomarkers may determine presymptomatic AD subjects and enable early intervention. Recently, Multitask sparse feature learning has been successfully applied to many computer vision and biomedical informatics researches. It aims to improve the generalization performance by exploiting the shared features among different tasks. However, most of the existing algorithms are formulated as a supervised learning scheme. Its drawback is with either insufficient feature numbers or missing label information. To address these challenges, we formulate an unsupervised framework for multi-task sparse feature learning based on a novel dictionary learning algorithm. To solve the unsupervised learning problem, we propose a two-stage Multi-Source Multi-Target Dictionary Learning (MMDL) algorithm. In stage 1, we propose a multi-source dictionary learning method to utilize the common and individual sparse features in different time slots. In stage 2, supported by a rigorous theoretical analysis, we develop a multi-task learning method to solve the missing label problem. Empirical studies on an N = 3970 longitudinal brain image data set, which involves 2 sources and 5 targets, demonstrate the improved prediction accuracy and speed efficiency of MMDL in comparison with other state-of-the-art algorithms.",
keywords = "Alzheimer’s disease, Dictionary learning, Multi-task",
author = "Jie Zhang and Qingyang Li and Caselli, {Richard John} and Thompson, {Paul M.} and Jieping Ye and Yalin Wang",
year = "2017",
doi = "10.1007/978-3-319-59050-9_15",
language = "English (US)",
isbn = "9783319590493",
volume = "10265 LNCS",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "184--197",
booktitle = "Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings",

}

TY - GEN

T1 - Multi-source multi-target dictionary learning for prediction of cognitive decline

AU - Zhang, Jie

AU - Li, Qingyang

AU - Caselli, Richard John

AU - Thompson, Paul M.

AU - Ye, Jieping

AU - Wang, Yalin

PY - 2017

Y1 - 2017

N2 - Alzheimer’s Disease (AD) is the most common type of dementia. Identifying correct biomarkers may determine presymptomatic AD subjects and enable early intervention. Recently, Multitask sparse feature learning has been successfully applied to many computer vision and biomedical informatics researches. It aims to improve the generalization performance by exploiting the shared features among different tasks. However, most of the existing algorithms are formulated as a supervised learning scheme. Its drawback is with either insufficient feature numbers or missing label information. To address these challenges, we formulate an unsupervised framework for multi-task sparse feature learning based on a novel dictionary learning algorithm. To solve the unsupervised learning problem, we propose a two-stage Multi-Source Multi-Target Dictionary Learning (MMDL) algorithm. In stage 1, we propose a multi-source dictionary learning method to utilize the common and individual sparse features in different time slots. In stage 2, supported by a rigorous theoretical analysis, we develop a multi-task learning method to solve the missing label problem. Empirical studies on an N = 3970 longitudinal brain image data set, which involves 2 sources and 5 targets, demonstrate the improved prediction accuracy and speed efficiency of MMDL in comparison with other state-of-the-art algorithms.

AB - Alzheimer’s Disease (AD) is the most common type of dementia. Identifying correct biomarkers may determine presymptomatic AD subjects and enable early intervention. Recently, Multitask sparse feature learning has been successfully applied to many computer vision and biomedical informatics researches. It aims to improve the generalization performance by exploiting the shared features among different tasks. However, most of the existing algorithms are formulated as a supervised learning scheme. Its drawback is with either insufficient feature numbers or missing label information. To address these challenges, we formulate an unsupervised framework for multi-task sparse feature learning based on a novel dictionary learning algorithm. To solve the unsupervised learning problem, we propose a two-stage Multi-Source Multi-Target Dictionary Learning (MMDL) algorithm. In stage 1, we propose a multi-source dictionary learning method to utilize the common and individual sparse features in different time slots. In stage 2, supported by a rigorous theoretical analysis, we develop a multi-task learning method to solve the missing label problem. Empirical studies on an N = 3970 longitudinal brain image data set, which involves 2 sources and 5 targets, demonstrate the improved prediction accuracy and speed efficiency of MMDL in comparison with other state-of-the-art algorithms.

KW - Alzheimer’s disease

KW - Dictionary learning

KW - Multi-task

UR - http://www.scopus.com/inward/record.url?scp=85020489007&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85020489007&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-59050-9_15

DO - 10.1007/978-3-319-59050-9_15

M3 - Conference contribution

SN - 9783319590493

VL - 10265 LNCS

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 184

EP - 197

BT - Information Processing in Medical Imaging - 25th International Conference, IPMI 2017, Proceedings

PB - Springer Verlag

ER -