Learning Under Delayed Feedback: Implicitly Adapting to Gradient Delays
Rotem Zamir Aviv 1 Ido Hakimi 2 Assaf Schuster 2 Kfir Y. Levy 1 3
Abstract ceed. Such methods are easier to analyze, yet their per-
We consider stochastic convex optimization formance depends on the slowest machine and they require
problems, where several machines act asyn- large communication overheads. Conversely, in the asyn-
chronously in parallel ...


雷达卡




京公网安备 11010802022788号







