Date of Award

2017

Document Type

Thesis

Degree Name

Bachelors

Department

Natural Sciences

First Advisor

McDonald, Patrick

Area of Concentration

Mathematics

Abstract

Brain images are a rich source of data with potential to improve the lives of millions. This data is usually private and protected; it cannot, in general, be freely shared. Given these constraints, implementations of algorithms used to process such data should not depend on pooling the data at a single location. For one important algorithm, Independent Vector Analysis (IVA), we develop a decentralized implementation that we call Decentralized data Independent Vector Analysis (DIVA). We study the performance of our algorithm and discuss the extent to which it provides a framework for the development of an implementation of IVA in which it is possible to give privacy guarantees.

Download

Share

COinS