Paper ID: 2410.09973
Gradient Span Algorithms Make Predictable Progress in High Dimension
Felix Benning, Leif Döring
We prove that all 'gradient span algorithms' have asymptotically deterministic behavior on scaled Gaussian random functions as the dimension tends to infinity. In particular, this result explains the counterintuitive phenomenon that different training runs of many large machine learning models result in approximately equal cost curves despite random initialization on a complicated non-convex landscape. The distributional assumption of (non-stationary) isotropic Gaussian random functions we use is sufficiently general to serve as realistic model for machine learning training but also encompass spin glasses and random quadratic functions.
Submitted: Oct 13, 2024