Breadcrumbs Lucerne University of Applied Sciences and Arts Lucerne School of Engineering and Architecture Campus Events Lecture «Learning Gaussian Mixtures via Method of Moments» Lecture «Learning Gaussian Mixtures via Method of Moments» Prof. Dr. Carlos Enriques Améndola Céron, of Technical University Munich, is guest at Lucerne School of Engineering and Architecture and is giving a public lecture about «Learning Gaussian Mixtures via Method of Moments». WED 27 November 2019 13:00 bis 14:00 Uhr place of event Hochschule Luzern – Technik & Architektur Technikumsstrasse 21, E526 6048 Horw Show on Google Maps Share Tweet LinkedIn E Mail Abstract: Following up on Karl Pearson’s introduction of the method of moments to recover the parameters of a mixture of two univariate Gaussians, we ask in general if the model parameters of a mixture of k Gaussians in dimension n can be identified from all the moments up to certain order d. We also study this identifiability problem in the submodel of homoscedastic mixtures, and provide a full answer when the common covariance matrix is known. Includes joint work with Daniele Agostini, Bernd Sturmfels and Kristian Ranestad.
