Loading lesson…
Adaptive Steps Per Coordinate
DetailedOverviewQuizFlashcard
Adaptive Steps Per Coordinate
g = 2, G = 4, η = 0.1: AdaGrad steps η/√G = 0.05. The dense neighbor at G = 100 steps 0.01; a shared G crushes both.
This Optimization lesson is part of a paid plan. Sign in with a plan that includes this course to read it.