An accelerated communication-efficient primal-dual optimization framework for structured machine learning

Chenxin Ma, Martin Jaggi, Frank E. Curtis, Nathan Srebro, Martin Takác