Model guided deep learning approach towards prediction of physical system behavior

Subhasish Das, Anurag Agrawal, Ayan Banerjee, Sandeep Gupta

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

1 Citation (Scopus)

Abstract

Cyber-physical control systems involve a discrete computational algorithm to control continuous physical systems. Often the control algorithm uses predictive models of the physical system in its decision making process. However, physical system models suffer from several inaccuracies when employed in practice. Mitigating such inaccuracies is often difficult and have to be repeated for different instances of the physical system. In this paper, we propose a model guided deep learning method for extraction of accurate prediction models of physical systems, in presence of artifacts observed in real life deployments. Given an initial potentially suboptimal mathematical prediction model, our model guided deep learning method iteratively improves the model through a data driven training approach. We apply the proposed approach on the closed loop blood glucose control system. Using this proposed approach, we achieve an improvement over predictive Bergman Minimal Model by a factor of around 100.

Original languageEnglish (US)
Title of host publicationProceedings - 16th IEEE International Conference on Machine Learning and Applications, ICMLA 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages1079-1082
Number of pages4
Volume2018-January
ISBN (Electronic)9781538614174
DOIs
StatePublished - Jan 16 2018
Event16th IEEE International Conference on Machine Learning and Applications, ICMLA 2017 - Cancun, Mexico
Duration: Dec 18 2017Dec 21 2017

Other

Other16th IEEE International Conference on Machine Learning and Applications, ICMLA 2017
CountryMexico
CityCancun
Period12/18/1712/21/17

Fingerprint

Control systems
Deep learning
Glucose
Blood
Decision making

Keywords

  • Artificial neural networks
  • deep learning
  • physical systems
  • prediction

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Science Applications

Cite this

Das, S., Agrawal, A., Banerjee, A., & Gupta, S. (2018). Model guided deep learning approach towards prediction of physical system behavior. In Proceedings - 16th IEEE International Conference on Machine Learning and Applications, ICMLA 2017 (Vol. 2018-January, pp. 1079-1082). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICMLA.2017.000-5

Model guided deep learning approach towards prediction of physical system behavior. / Das, Subhasish; Agrawal, Anurag; Banerjee, Ayan; Gupta, Sandeep.

Proceedings - 16th IEEE International Conference on Machine Learning and Applications, ICMLA 2017. Vol. 2018-January Institute of Electrical and Electronics Engineers Inc., 2018. p. 1079-1082.

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

Das, S, Agrawal, A, Banerjee, A & Gupta, S 2018, Model guided deep learning approach towards prediction of physical system behavior. in Proceedings - 16th IEEE International Conference on Machine Learning and Applications, ICMLA 2017. vol. 2018-January, Institute of Electrical and Electronics Engineers Inc., pp. 1079-1082, 16th IEEE International Conference on Machine Learning and Applications, ICMLA 2017, Cancun, Mexico, 12/18/17. https://doi.org/10.1109/ICMLA.2017.000-5
Das S, Agrawal A, Banerjee A, Gupta S. Model guided deep learning approach towards prediction of physical system behavior. In Proceedings - 16th IEEE International Conference on Machine Learning and Applications, ICMLA 2017. Vol. 2018-January. Institute of Electrical and Electronics Engineers Inc. 2018. p. 1079-1082 https://doi.org/10.1109/ICMLA.2017.000-5
Das, Subhasish ; Agrawal, Anurag ; Banerjee, Ayan ; Gupta, Sandeep. / Model guided deep learning approach towards prediction of physical system behavior. Proceedings - 16th IEEE International Conference on Machine Learning and Applications, ICMLA 2017. Vol. 2018-January Institute of Electrical and Electronics Engineers Inc., 2018. pp. 1079-1082
@inproceedings{449c93a124c348818c5b2c4fa49523ee,
title = "Model guided deep learning approach towards prediction of physical system behavior",
abstract = "Cyber-physical control systems involve a discrete computational algorithm to control continuous physical systems. Often the control algorithm uses predictive models of the physical system in its decision making process. However, physical system models suffer from several inaccuracies when employed in practice. Mitigating such inaccuracies is often difficult and have to be repeated for different instances of the physical system. In this paper, we propose a model guided deep learning method for extraction of accurate prediction models of physical systems, in presence of artifacts observed in real life deployments. Given an initial potentially suboptimal mathematical prediction model, our model guided deep learning method iteratively improves the model through a data driven training approach. We apply the proposed approach on the closed loop blood glucose control system. Using this proposed approach, we achieve an improvement over predictive Bergman Minimal Model by a factor of around 100.",
keywords = "Artificial neural networks, deep learning, physical systems, prediction",
author = "Subhasish Das and Anurag Agrawal and Ayan Banerjee and Sandeep Gupta",
year = "2018",
month = "1",
day = "16",
doi = "10.1109/ICMLA.2017.000-5",
language = "English (US)",
volume = "2018-January",
pages = "1079--1082",
booktitle = "Proceedings - 16th IEEE International Conference on Machine Learning and Applications, ICMLA 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Model guided deep learning approach towards prediction of physical system behavior

AU - Das, Subhasish

AU - Agrawal, Anurag

AU - Banerjee, Ayan

AU - Gupta, Sandeep

PY - 2018/1/16

Y1 - 2018/1/16

N2 - Cyber-physical control systems involve a discrete computational algorithm to control continuous physical systems. Often the control algorithm uses predictive models of the physical system in its decision making process. However, physical system models suffer from several inaccuracies when employed in practice. Mitigating such inaccuracies is often difficult and have to be repeated for different instances of the physical system. In this paper, we propose a model guided deep learning method for extraction of accurate prediction models of physical systems, in presence of artifacts observed in real life deployments. Given an initial potentially suboptimal mathematical prediction model, our model guided deep learning method iteratively improves the model through a data driven training approach. We apply the proposed approach on the closed loop blood glucose control system. Using this proposed approach, we achieve an improvement over predictive Bergman Minimal Model by a factor of around 100.

AB - Cyber-physical control systems involve a discrete computational algorithm to control continuous physical systems. Often the control algorithm uses predictive models of the physical system in its decision making process. However, physical system models suffer from several inaccuracies when employed in practice. Mitigating such inaccuracies is often difficult and have to be repeated for different instances of the physical system. In this paper, we propose a model guided deep learning method for extraction of accurate prediction models of physical systems, in presence of artifacts observed in real life deployments. Given an initial potentially suboptimal mathematical prediction model, our model guided deep learning method iteratively improves the model through a data driven training approach. We apply the proposed approach on the closed loop blood glucose control system. Using this proposed approach, we achieve an improvement over predictive Bergman Minimal Model by a factor of around 100.

KW - Artificial neural networks

KW - deep learning

KW - physical systems

KW - prediction

UR - http://www.scopus.com/inward/record.url?scp=85048475479&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048475479&partnerID=8YFLogxK

U2 - 10.1109/ICMLA.2017.000-5

DO - 10.1109/ICMLA.2017.000-5

M3 - Conference contribution

AN - SCOPUS:85048475479

VL - 2018-January

SP - 1079

EP - 1082

BT - Proceedings - 16th IEEE International Conference on Machine Learning and Applications, ICMLA 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -