Gradient Mechanism to Preserve Differential Privacy and Deter Against Model Inversion Attacks in Healthcare Analytics

Alexander Krall, Daniel Antion Finke, Hui Yang

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

Abstract

Advanced sensing technologies, driven by the Internet of Things, have caused a sharp increase in data availability within the healthcare system. The newfound availability of data offers an unprecedented opportunity to develop new analytical methods to improve the quality of patient care. Data availability, however, is a double-edged sword. Malicious attacks and data breaches are increasingly seen in the healthcare field, which result in costly disruptions to operations. Adversaries exploit analytic models to infer participation in a dataset or estimate sensitivity attributes about a target patient. This paper is aimed at developing a differentially private gradient-based mechanism and assessing its utility in mitigating the impact of these attack risks within the context of the intensive care units. Experimental results showed that this methodology is capable of greatly reducing the risk of model inversion while retaining model accuracy. Thus, health systems that employ this technique can be given more peace of mind that high-quality services can be delivered in such a way that privacy is preserved.

Original languageEnglish (US)
Title of host publication42nd Annual International Conferences of the IEEE Engineering in Medicine and Biology Society
Subtitle of host publicationEnabling Innovative Technologies for Global Healthcare, EMBC 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages5714-5717
Number of pages4
ISBN (Electronic)9781728119908
DOIs
StatePublished - Jul 2020
Event42nd Annual International Conferences of the IEEE Engineering in Medicine and Biology Society, EMBC 2020 - Montreal, Canada
Duration: Jul 20 2020Jul 24 2020

Publication series

NameProceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS
Volume2020-July
ISSN (Print)1557-170X

Conference

Conference42nd Annual International Conferences of the IEEE Engineering in Medicine and Biology Society, EMBC 2020
CountryCanada
CityMontreal
Period7/20/207/24/20

All Science Journal Classification (ASJC) codes

  • Signal Processing
  • Biomedical Engineering
  • Computer Vision and Pattern Recognition
  • Health Informatics

Fingerprint Dive into the research topics of 'Gradient Mechanism to Preserve Differential Privacy and Deter Against Model Inversion Attacks in Healthcare Analytics'. Together they form a unique fingerprint.

Cite this