Differentially Private Release of Spatio-Temporal Data Statistics

Explore a technical lecture on differential privacy in spatio-temporal data analysis presented by Dr. V. Arvind Rameshwar, Research Fellow at India Urban Data Exchange. Delve into strategies for implementing user-level differential privacy (DP) when releasing sample mean and variance statistics from spatio-temporal datasets, with emphasis on minimizing reconstruction errors. Learn about a novel computable metric for measuring worst-case reconstruction error and discover an error-preserving algorithm that enhances privacy loss through the composition of user-level DP mechanisms across disjoint spatio-temporal regions. Examine real-world applications through case studies using Intelligent Traffic Management System (ITMS) data from an Indian city. Benefit from Dr. Rameshwar's expertise, developed through his distinguished academic career at BITS Pilani University and Indian Institute of Science, where his research contributions earned multiple prestigious awards including the Prime Minister's Research Fellowship and Qualcomm Innovation Fellowships.