44 citations · 174 across the 35 of their papers we have counts for
Showing 2018 · cs.LGShow all
2 papers · 2 filters
cs.LG2018
Online Learning: Sufficient Statistics and the Burkholder Method
Dylan J. Foster, Alexander Rakhlin, Karthik Sridharan
We uncover a fairly general principle in online learning: If regret can be (approximately) expressed as a function of certain "sufficient statistics" for the data sequence, then th…
cs.LG2018★ 44 cited
Theory of Deep Learning IIb: Optimization Properties of SGD
Chiyuan Zhang, Qianli Liao, Alexander Rakhlin +3
In Theory IIb we characterize with a mix of theory and experiments the optimization of deep convolutional networks by Stochastic Gradient Descent. The main new result in this paper…