Enabling Users with Ease of Application of Different Machine Learning Models and Verification of Results; Consumer Empowerment in Digital Privacy in CCPA and GDPR

For my STS research project, I studied digital privacy and compared California Consumer Privacy Act (CCPA) and General Data Protection Regulation (GDPR) with respect to consumer empowerment by conducting literature review on scholarships that analyze these two documents. I was interested in digital privacy because it had been observably difficult to maintain control over one’s personal information online. My key objectives were to learn more about the practices and regulation of digital privacy and find out what action one can take proactively as a consumer to protect one’s personal information online.
I started out to find analysis about CCPA and GDPR on the macro level, such as what attitude each policy takes towards preserving consumer privacy and what stakeholders and incentives are behind each policy. As I researched, I revised my goal to be studying what each policy implements that empowers consumers and how effective such implementations are. My findings are that while CCPA and GDPR are both forward-looking regulations in digital privacy, their enforceability is very limited, and the progress on consumer empowerment are far from satisfactory. The only question left unanswered was why the progress is so limited as a common pattern when regulations on digital privacy are desired by the people and could bring benefit to all individual consumers.
For my technical project, I designed a web application that enables users to verify the results of available machine learning models easily and conveniently. The papers in the field of machine learning are often published without source code given copyright and plagiarism concerns. The lack of source code and the variability of hyperparameters in machine learning models makes the reproduction of experiment results very difficult. To incentivize researchers to publish a portion of their implementation, I proposed this web application to allow authors of papers to enable other researchers to verify results of exported machine learning models without making the models public to the internet. This web application was designed with an internet-scalable architecture for high-volume traffic and fine-grained access control for security and privacy. Future work could explore the potential to incorporate more cloud computing components to allow customizable input preprocessing and faster content delivery.

School of Engineering and Applied Science
Bachelor of Science in Computer Science
Technical Advisor: Upsorn Praphamontripong
STS Advisor: Sean Ferguson
Technical Team Members: Yu Yang