Modern large-scale machine learning models are often trained by the parallel stochastic gradient descent algorithm or its variants on distributed systems with parameter servers or master/worker nodes. However, the communications of the gradient aggregation and model synchronization between the master and worker nodes are the major obstacles for ...
Creator:
Yan, Ming (Michigan State University)
Created:
2019-10-14
Contributed By:
University of Minnesota, Institute for Mathematics and its Applications.
There are several primal-dual algorithms for minimizing f(x)+g(x)+h(Ax), where f, g, and h are convex functions, f is differentiable with a Lipschitz continuous gradient, and A is a bounded linear operator. In this talk, I will introduce several different primal-dual algorithms for minimizing the sum of three functions and their connections with...
Creator:
Yan, Ming (Michigan State University)
Created:
2018-10-26
Contributed By:
University of Minnesota, Institute for Mathematics and its Applications.
