Partly Supervised Multi-Task Learning

Abdullah Al Zubaer Imran; Chao Huang; Hui Tang; Wei Fan; Yuan Xiao; Dingjun Hao; Zhen Qian; Demetri Terzopoulos

doi:10.1109/ICMLA51294.2020.00126

Partly Supervised Multi-Task Learning

Abdullah Al Zubaer Imran, Chao Huang, Hui Tang, Wei Fan, Yuan Xiao, Dingjun Hao, Zhen Qian, Demetri Terzopoulos

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

7 Scopus citations

Abstract

Semi-supervised learning has recently been attracting attention as an alternative to fully supervised models that require large pools of labeled data. Moreover, optimizing a model for multiple tasks can provide better generalizability than single-task learning. Leveraging self-supervision and adversarial training, we propose a novel, general purpose semi-supervised, multiple-task model - namely, self-supervised, semi-supervised, multi-task learning (S4MTL) - for accomplishing two important medical image analysis tasks: segmentation and diagnostic classification. Experimental results on chest and spine X-ray datasets confirm that our S4MTL model significantly outperforms semi-supervised single-task, semi/fully-supervised multi-task, and fully-supervised single-task models, even with a 50% reduction in class and segmentation labels.

Original language	English
Title of host publication	Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020
Editors	M. Arif Wani, Feng Luo, Xiaolin Li, Dejing Dou, Francesco Bonchi
Pages	769-774
Number of pages	6
ISBN (Electronic)	9781728184708
DOIs	https://doi.org/10.1109/ICMLA51294.2020.00126
State	Published - Dec 2020
Event	19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020 - Virtual, Miami, United States Duration: Dec 14 2020 → Dec 17 2020

Publication series

Name	Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020

Conference

Conference	19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020
Country/Territory	United States
City	Virtual, Miami
Period	12/14/20 → 12/17/20

Bibliographical note

Publisher Copyright:
© 2020 IEEE.

Keywords

X-ray
chest
classification
multi-task learning
segmentation
self-supervision
semi-supervised learning
spine

ASJC Scopus subject areas

Artificial Intelligence
Computer Science Applications
Computer Vision and Pattern Recognition
Hardware and Architecture

Access to Document

10.1109/ICMLA51294.2020.00126

Cite this

Imran, A. A. Z., Huang, C., Tang, H., Fan, W., Xiao, Y., Hao, D., Qian, Z., & Terzopoulos, D. (2020). Partly Supervised Multi-Task Learning. In M. A. Wani, F. Luo, X. Li, D. Dou, & F. Bonchi (Eds.), Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020 (pp. 769-774). Article 9356271 (Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020). https://doi.org/10.1109/ICMLA51294.2020.00126

Imran, Abdullah Al Zubaer ; Huang, Chao ; Tang, Hui et al. / Partly Supervised Multi-Task Learning. Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020. editor / M. Arif Wani ; Feng Luo ; Xiaolin Li ; Dejing Dou ; Francesco Bonchi. 2020. pp. 769-774 (Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020).

@inproceedings{49e96a59a4794c19bff0b1fdeaf7841b,

title = "Partly Supervised Multi-Task Learning",

abstract = "Semi-supervised learning has recently been attracting attention as an alternative to fully supervised models that require large pools of labeled data. Moreover, optimizing a model for multiple tasks can provide better generalizability than single-task learning. Leveraging self-supervision and adversarial training, we propose a novel, general purpose semi-supervised, multiple-task model - namely, self-supervised, semi-supervised, multi-task learning (S4MTL) - for accomplishing two important medical image analysis tasks: segmentation and diagnostic classification. Experimental results on chest and spine X-ray datasets confirm that our S4MTL model significantly outperforms semi-supervised single-task, semi/fully-supervised multi-task, and fully-supervised single-task models, even with a 50% reduction in class and segmentation labels.",

keywords = "X-ray, chest, classification, multi-task learning, segmentation, self-supervision, semi-supervised learning, spine",

author = "Imran, {Abdullah Al Zubaer} and Chao Huang and Hui Tang and Wei Fan and Yuan Xiao and Dingjun Hao and Zhen Qian and Demetri Terzopoulos",

note = "Publisher Copyright: {\textcopyright} 2020 IEEE.; 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020 ; Conference date: 14-12-2020 Through 17-12-2020",

year = "2020",

month = dec,

doi = "10.1109/ICMLA51294.2020.00126",

language = "English",

series = "Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020",

pages = "769--774",

editor = "Wani, {M. Arif} and Feng Luo and Xiaolin Li and Dejing Dou and Francesco Bonchi",

booktitle = "Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020",

}

Imran, AAZ, Huang, C, Tang, H, Fan, W, Xiao, Y, Hao, D, Qian, Z & Terzopoulos, D 2020, Partly Supervised Multi-Task Learning. in MA Wani, F Luo, X Li, D Dou & F Bonchi (eds), Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020., 9356271, Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020, pp. 769-774, 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020, Virtual, Miami, United States, 12/14/20. https://doi.org/10.1109/ICMLA51294.2020.00126

Partly Supervised Multi-Task Learning. / Imran, Abdullah Al Zubaer; Huang, Chao; Tang, Hui et al.
Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020. ed. / M. Arif Wani; Feng Luo; Xiaolin Li; Dejing Dou; Francesco Bonchi. 2020. p. 769-774 9356271 (Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Partly Supervised Multi-Task Learning

AU - Imran, Abdullah Al Zubaer

AU - Huang, Chao

AU - Tang, Hui

AU - Fan, Wei

AU - Xiao, Yuan

AU - Hao, Dingjun

AU - Qian, Zhen

AU - Terzopoulos, Demetri

PY - 2020/12

Y1 - 2020/12

N2 - Semi-supervised learning has recently been attracting attention as an alternative to fully supervised models that require large pools of labeled data. Moreover, optimizing a model for multiple tasks can provide better generalizability than single-task learning. Leveraging self-supervision and adversarial training, we propose a novel, general purpose semi-supervised, multiple-task model - namely, self-supervised, semi-supervised, multi-task learning (S4MTL) - for accomplishing two important medical image analysis tasks: segmentation and diagnostic classification. Experimental results on chest and spine X-ray datasets confirm that our S4MTL model significantly outperforms semi-supervised single-task, semi/fully-supervised multi-task, and fully-supervised single-task models, even with a 50% reduction in class and segmentation labels.

AB - Semi-supervised learning has recently been attracting attention as an alternative to fully supervised models that require large pools of labeled data. Moreover, optimizing a model for multiple tasks can provide better generalizability than single-task learning. Leveraging self-supervision and adversarial training, we propose a novel, general purpose semi-supervised, multiple-task model - namely, self-supervised, semi-supervised, multi-task learning (S4MTL) - for accomplishing two important medical image analysis tasks: segmentation and diagnostic classification. Experimental results on chest and spine X-ray datasets confirm that our S4MTL model significantly outperforms semi-supervised single-task, semi/fully-supervised multi-task, and fully-supervised single-task models, even with a 50% reduction in class and segmentation labels.

KW - X-ray

KW - chest

KW - classification

KW - multi-task learning

KW - segmentation

KW - self-supervision

KW - semi-supervised learning

KW - spine

UR - http://www.scopus.com/inward/record.url?scp=85101697387&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85101697387&partnerID=8YFLogxK

U2 - 10.1109/ICMLA51294.2020.00126

DO - 10.1109/ICMLA51294.2020.00126

M3 - Conference contribution

AN - SCOPUS:85101697387

T3 - Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020

SP - 769

EP - 774

BT - Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020

A2 - Wani, M. Arif

A2 - Luo, Feng

A2 - Li, Xiaolin

A2 - Dou, Dejing

A2 - Bonchi, Francesco

T2 - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020

Y2 - 14 December 2020 through 17 December 2020

ER -

Imran AAZ, Huang C, Tang H, Fan W, Xiao Y, Hao D et al. Partly Supervised Multi-Task Learning. In Wani MA, Luo F, Li X, Dou D, Bonchi F, editors, Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020. 2020. p. 769-774. 9356271. (Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020). doi: 10.1109/ICMLA51294.2020.00126

Partly Supervised Multi-Task Learning

Abstract

Publication series

Conference

Bibliographical note

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this