Getting More from Less: Optimal Estimation and Learning, For Sparse, High Dimensional, or Untrusted Data.
Abstract
The goals of the proposed research are to develop computationally efficient and information theo-retically optimal algorithms and estimators for a variety of fundamental problems. The problems we focus on center on two theoretically rich and practically important" settings: extracting accurate information from complex distributions given relatively sparse samples, and obtaining accurate estima"tion and learning algorithms that can be applied to datasets where some (unknown) portion of the data is untrusted.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Jun 09, 2017
- Source ID
- N000141712562
Entities
People
- Gregory Valiant
Organizations
- Office of Naval Research
- Stanford University
- United States Navy