"Making Gradient Descent Optimal for Strongly Convex Stochastic Optimization."

Alexander Rakhlin, Ohad Shamir, Karthik Sridharan (2012)
a service of Schloss Dagstuhl - Leibniz Center for Informatics