Viterbi Faculty Directory

Research Summary

---

Developed differentially private algorithms for common data mining tasks such as search log release and malware detection in the local model of privacy, leading to the first commercial deployment of differential privacy.

Identified privacy shortcomings in microtargeted advertising, Apple's deployment of differential privacy, and treatment of Bluetooth permissions by mobile OSes, leading to changes in practices and greater transparency.

Demonstrated that Facebook's ad delivery optimization algorithms can direct job and housing ads to audiences skewed by race & gender even when advertisers target large, inclusive audiences. Furthermore, showed that Facebook's ad delivery optimization algorithms limit political advertisers' ability to reach audiences that do not share their political views and can create informational filter bubbles. Finally, developed a new auditing methodology to distinguish between skew introduced by job ad delivery algorithms that may be explainable by differences in qualifications (permissible by law) from skew due to factors such as the ad platform's optimization for its business objectives, demonstrating that Facebook's job ad delivery algorithms are not merely biased but discriminatory.

Showed how automated analyses of users' privacy-related actions can be used to improve product privacy features.