Governance for Big Data - July 2018

PI(s), Co-PI(s), Researchers: Deirdre Mulligan (U.C. Berkeley)

HARD PROBLEM(S) ADDRESSED
Human Behavior, Policy-Governed Secure Collaboration

PUBLICATIONS

KEY HIGHLIGHTS

We have begun mapping the existing literature, especially technical literature, on privacy by design, provable privacy, and value-sensitive engineering. We have also done initial work on a project to measure how likely it is that data analysis will reach unacceptable models naturally to understand when frameworks must overcome problems vs. when known workflows will lead to acceptable outcomes. Finally, we have begun a project comparing existing toolkits for measuring fairness in classification models that examines how these tools can be built into common data science workflows.

COMMUNITY ENGAGEMENTS

We are planning workshops for stakeholders, which we expect to conduct over the coming year.

EDUCATIONAL ADVANCES: