Pursuit of Low-dimensional Structures in High-dimensional Data - 2013

Explore cutting-edge techniques for modeling and extracting low-dimensional structures from high-dimensional data in this lecture by Yi Ma from Microsoft Research. Delve into a new class of models that effectively handle nonlinear transformations, gross corruption, and compressed measurements in images and videos. Learn about recent advancements in convex optimization for recovering low-rank or sparse signals, offering both strong theoretical guarantees and efficient algorithms for solving complex high-dimensional problems. Discover how these mathematical models and tools can revolutionize computer vision, image processing, and pattern recognition tasks. Examine emerging applications of these techniques to diverse data types, including web documents, image tags, microarray data, audio/music analysis, and graphical models. Gain insights from Yi Ma's extensive research experience and collaborations with leading institutions in the field of visual computing and high-dimensional data analysis.