1 paper
Chien-Chih Wang, Kent Loong Tan, Chih-Jen Lin
Deep learning involves a difficult non-convex optimization problem, which is often solved by stochastic gradient (SG) methods. While SG is usually effective, it may not be robust i…