Learning Mixed Membership Community Models- A statistical and a Computational Framework
Abstract
Developed tensor sketching methods for efficient dimensionality reduction in large datasets. Analyzed and derived guarantees of tensor sketches and the advantages over previous methods. Showed gains in deep learning applications using these sketching techniques.
Document Details
- Document Type
- Technical Report
- Publication Date
- Nov 22, 2019
- Accession Number
- AD1097125
Entities
People
- Anima Anandkumar
Organizations
- University of California