3 papers
cs.LG2026
A prism hierarchy of learning regimes in large linear autoencoders
Eugene Golikov, Yaroslav Gusev, Dmitry Yarotsky
Theoretical studies of machine learning models commonly consider different limiting regimes in which the learning dynamics of gradient descent becomes theoretically tractable. It i…
math.OC2025
Corner Gradient Descent
Dmitry Yarotsky
We consider SGD-type optimization on infinite-dimensional quadratic problems with power law spectral conditions. It is well-known that on such problems deterministic GD has loss co…
cs.LG2024
SGD with memory: fundamental properties and stochastic acceleration
Dmitry Yarotsky, Maksim Velikanov
An important open problem is the theoretically feasible acceleration of mini-batch SGD-type algorithms on quadratic problems with power-law spectrum. In the non-stochastic setting,…