Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Explore a cutting-edge approach to differentially private estimation in this Google TechTalk presented by Gavin Brown. Learn about innovative algorithms that overcome significant limitations of existing methods, including the need for strong prior bounds on data and poor performance in high-dimensional settings. Discover how a novel stabilized greedy outlier-removal process enables fast and near-optimal private algorithms for fundamental statistical tasks including mean estimation, covariance estimation, and least squares regression. Understand the theoretical foundations and practical implications of this research that addresses the exponential time complexity issues plaguing many current differential privacy algorithms. Gain insights into collaborative work spanning multiple institutions and access to the underlying research through provided arXiv links for deeper technical exploration.
