Mean Estimation in High-Dimensional Binary Markov Gaussian Mixture Models
HUJI Machine Learning Club via YouTube
Overview
Save Big on Coursera Plus. 7,000+ courses at $160 off. Limited Time Only!
Explore a 59-minute lecture on mean estimation in high-dimensional binary Markov Gaussian mixture models, delivered by Nir Weinberger from Technion. Delve into the challenges of estimating parameters when observing samples from two distinct populations with latent identities, particularly when adjacent samples are likely drawn from the same population. Learn about optimal minimax rates for estimation error and understand how they bridge the gap between standard Gaussian location models and symmetric Gaussian mixture models. Discover the complex relationships between data memory, sample size, dimension, and signal strength in these statistical problems. Based on joint research with Yihan Zhang from IST Austria, gain insights from Weinberger, an assistant Professor at the Technion's Viterbi Faculty of Electrical and Computer Engineering, who brings extensive experience in communications and signal processing.
Syllabus
Delivered on Thursday, December 15th, 2022, AM
Taught by
HUJI Machine Learning Club