Dive into the world of Independent Component Analysis (ICA) through this comprehensive lecture series. Explore the fundamental assumption that data is a mixture of statistically independent signals and learn how to recover original signals without prior knowledge of the mixing process. Discover the standard case of linear mixture with equal numbers of data components and signals, while also touching on potential extensions. Gain insights into various ICA approaches, with a focus on a cumulant-based method. Progress through six detailed lectures, starting with an overview and gradually delving into more complex aspects of ICA, including its applications in blind source separation. Master key concepts and techniques over the course of 2 hours and 30 minutes, equipping yourself with valuable knowledge in this important area of signal processing and data analysis.
Overview
Syllabus
ICA 0 - ICA in a Nutshell (23 min).
ICA 1.1a - Independent Component Analysis (21 min).
ICA 1.1b-1.5 - Independent Component Analysis (34 min).
ICA 2.1-2.3 - Independent Component Analysis (29 min).
ICA 2.4-2.6 - Independent Component Analysis (20 min).
ICA 2.7-2.8 - Independent Component Analysis (10 min).
Taught by
Prof. Laurenz Wiskott
