Yue Lu | Nov 30, 2021 | Learning by Random Features and Kernel Random Matrices

Описание: Title: Learning by Random Features and Kernel Random Matrices: Sharp Asymptotics and Universality Laws

Speaker: Yue Lu

Abstract: : Many new random matrix ensembles arise in learning and modern signal processing. The spectral properties of these matrices help answer crucial questions regarding the training and generalization performance of neural networks, and the fundamental limits of high-dimensional signal
recovery. As a result, there has been growing interest in precisely understanding the spectra and other asymptotic properties of these matrices. These new random matrices are often highly structured and
are the result of nonlinear transformations. This combination of structure and nonlinearity leads to substantial technical challenges when applying existing tools from random matrix theory to these new random matrix ensembles.