A Geometric View of Optimal Transport and Generative Models - ICBS 2024
Offered By: BIMSA via YouTube
Course Description
Overview
Explore a geometric interpretation of optimal transport theory and its application to generative models in this 52-minute lecture. Delve into the principles behind text-to-video generative models like Sora, analyzing their successes and limitations. Examine the physical errors in generated videos, including contradictions between correlation and causality, local coherence versus global absurdity, and the absence of critical states. Gain insights into the intrinsic reasons for these mistakes and discover proposed methods for improving generative models through the lens of optimal transport theory.
Syllabus
Xianfeng Gu: A Geometric View of Optimal Transport and Generative Models #ICBS2024
Taught by
BIMSA
Related Courses
Optimal Transport and PDE - Gradient Flows in the Wasserstein MetricSimons Institute via YouTube Crash Course on Optimal Transport
Simons Institute via YouTube Learning From Ranks, Learning to Rank - Jean-Philippe Vert, Google Brain
Alan Turing Institute via YouTube Optimal Transport for Machine Learning - Gabriel Peyre, Ecole Normale Superieure
Alan Turing Institute via YouTube Regularization for Optimal Transport and Dynamic Time Warping Distances - Marco Cuturi
Alan Turing Institute via YouTube