1.6k citations · 3.9k across the 64 of their papers we have counts for
Showing 2017Show all
3 papers · 1 filter
cs.LG2017★ 120 cited
The Power of Interpolation: Understanding the Effectiveness of SGD in Modern Over-parametrized Learning
Siyuan Ma, Raef Bassily, Mikhail Belkin
In this paper we aim to formally explain the phenomenon of fast convergence of SGD observed in modern machine learning. The key observation is that most modern learning architectur…
stat.ML2017★ 32 cited
Unperturbed: spectral analysis beyond Davis-Kahan
Justin Eldridge, Mikhail Belkin, Yusu Wang
Classical matrix perturbation results, such as Weyl's theorem for eigenvalues and the Davis-Kahan theorem for eigenvectors, are general purpose. These classical bounds are tight in…
stat.ML2017★ 31 cited
Diving into the shallows: a computational perspective on large-scale shallow learning
Siyuan Ma, Mikhail Belkin
In this paper we first identify a basic limitation in gradient descent-based optimization methods when used in conjunctions with smooth kernels. An analysis based on the spectral p…