Skip to contentSkip to site navigation
Please visit our COVID-19 (novel coronavirus) Updates page for all developments.
Computer Science
2020 Project Proposal

Decentralized Policy-Based Private Analytics

Jason Waterman (Computer Science)            

Ubiquitous sensing and data collection through our personal devices and our environment have created the potential to offer innovative services and applications while creating significant value from our personal data. These data are often used for a variety of personal and public Machine Learning (ML) models, such as building targeted advertising, content recommendation, and health analytics.  Building these models present several challenges: trust, as the user requires guarantees that their provided data will not be misused by the service provider; security, as these rich data resources are valuable to criminals; and scalability, as the costs of building and maintaining large data-centers can be prohibitive.

To tackle these challenges, we are building a scalable, privacy preserving approach to personal data processing. Rather than centrally collecting data from a user population for later processing, we move the required computation to the network edge where data logically resides and can be verified to be used in accordance with a user's privacy policy. This model is a better fit for current privacy and data processing regulation and allows data subjects to bear more of the cost burden of providing storage, computation and connectivity while preserving privacy.

Strong Python programming skills are a must.  In addition, students should have already taken CMPU-203 and CMPU-224.

How should students express interest in this project?
Students interested in this project should email me to set up a meeting to discuss the project.

This is a 10 week project running from May 27-July 31