Security and Privacy Preserving Deep Learning Framework that Protect Healthcare Data Breaches

Authors

  • S. Sreeji PG Scholar, Department of Computer Science and Engineering, Nehru College of Engineering and Research Centre, Palakkad, India
  • S. Shiji Assistant Professor, Department of Computer Science and Engineering, Nehru College of Engineering and Research Centre, Palakkad, India
  • M. Vysagh Assistant Professor, Department of Computer Science and Engineering, Nehru College of Engineering and Research Centre, Palakkad, India
  • T. Ambikadevi Amma Professor & Principal, Department of Computer Science and Engineering, Nehru College of Engineering and Research Centre, Palakkad, India

Keywords:

Anonymization, Bigdata, Back Propagation, Data Security, Gradient, SplitNN, Synchronous Optimization

Abstract

Big healthcare data security and privacy are a big concern increasing year-by-year. Heterogeneous data called big data, plays overwhelming role in medical industry. More than 750 data breaches occurred in 2015.The top data security breaches occurred from health care industry. The most important data security issue occurs during sharing sensitive data to train the system. There are several methods to protect the privacy of such healthcare data. Among them a distributed deep learning method called SplitNN, is the one which does not share raw data or model details with collaborating institutions (hospitals). Another method is sequentially sharing models in cyclic in order to train deep neural networks. Another approach is synchronous optimization approach which is empirically validated and shown to converge faster and to better test accuracies. The existing systems uses anonymization techniques to protect the privacy. The proposed deep learning framework keep patient’s original data in local platforms and send gradient values to the client and back propagate the data without any anonymization. The learning performance improves by using data from different platforms (hospitals) during training.

Downloads

Download data is not yet available.

Downloads

Published

20-07-2020

How to Cite

[1]
S. . Sreeji, S. . Shiji, M. . Vysagh, and T. A. . Amma, “Security and Privacy Preserving Deep Learning Framework that Protect Healthcare Data Breaches”, IJRESM, vol. 3, no. 7, pp. 148–152, Jul. 2020.

Issue

Section

Articles