Anytime MiniBatch: Exploiting Stragglers in Online Distributed Optimization
This addresses a bottleneck in large-scale machine learning by improving efficiency in distributed systems, though it is an incremental method building on existing optimization techniques.
The paper tackles the problem of stragglers slowing distributed optimization by proposing Anytime Minibatch, which allows nodes to compute gradients for variable minibatch sizes within fixed time limits, resulting in up to 1.5 times faster performance in Amazon EC2 and up to five times faster with high compute variability.
Distributed optimization is vital in solving large-scale machine learning problems. A widely-shared feature of distributed optimization techniques is the requirement that all nodes complete their assigned tasks in each computational epoch before the system can proceed to the next epoch. In such settings, slow nodes, called stragglers, can greatly slow progress. To mitigate the impact of stragglers, we propose an online distributed optimization method called Anytime Minibatch. In this approach, all nodes are given a fixed time to compute the gradients of as many data samples as possible. The result is a variable per-node minibatch size. Workers then get a fixed communication time to average their minibatch gradients via several rounds of consensus, which are then used to update primal variables via dual averaging. Anytime Minibatch prevents stragglers from holding up the system without wasting the work that stragglers can complete. We present a convergence analysis and analyze the wall time performance. Our numerical results show that our approach is up to 1.5 times faster in Amazon EC2 and it is up to five times faster when there is greater variability in compute node performance.