The Matching Principle: A Geometric Theory of Loss Functions for Nuisance-Robust Representation Learning
by Vishal Rajput
View on Kurate.org