Spectral Learning for High-Dimensional Tensors

• Activity Number: 87 - Reduced-Rank Methods: Seventy Years of History and State-of-the-Art Developments
• Type: Invited
• Date/Time: Monday, August 8, 2022 : 8:30 AM to 10:20 AM
• Sponsor: General Methodology
• Author(s): Ming Yuan*
• Companies: Columbia University
• Keywords: perturbation analysis; singular value decomposition; tensor

Abstract

Many problems can be formulated as recovering a spectral learning of tensors. Although an increasingly common task, it remains a challenging problem because of the lack of tools for perturbation analysis similar to those for matrices. In this talk, I will describe some recent results towards a general treatment of these problems.