ALGEBRAIC AND GEOMETRIC METHODS FOR CLUSTERING AND CLASSIFICATION
Abstract
This research will consider several computational problems that the PI is tackling. Many of these problems share the characteristic that data of interest are “hiding”, not only because of very low SNR, but, more challengingly, by their very nature as low-dimensional phenomena (either linear or nonlinear) observed in very large dimensional settings and obscured by various transformations. In particular, the PI proposes new mathematical frameworks for the important data discovery tasks of clustering and classification that combine modern techniques ranging from deep convolutional networks to high dimensional statistics, representation theory, information theory, and harmonic analysis.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 12, 2021
- Source ID
- FA95502010266
Entities
People
- Amit Singer
Organizations
- Air Force Office of Scientific Research
- Trustees of Princeton University
- United States Air Force