We propose Batch-Expansion Training (BET), a framework for running a batch optimizer on a gradually expanding dataset. As opposed to stochastic approaches, batches do not need to be resampled i.i.d. at every iteration, thus making BET more resource efficient in a distributed setting, and when disk-access is constrained. Moreover, BET can be easily paired with most batch optimizers, does not require any parameter-tuning, and compares favorably to existing stochastic and batch methods. We show that when the batch size grows exponentially with the number of outer iterations, BET achieves optimal O(1/ϵ) data-access convergence rate for strongly convex objectives. Experiments in parallel and distributed settings show that BET performs better than standard batch and stochastic approaches.
@article{arxiv.1704.06731,
title = {Batch-Expansion Training: An Efficient Optimization Framework},
author = {Michał Dereziński and Dhruv Mahajan and S. Sathiya Keerthi and S. V. N. Vishwanathan and Markus Weimer},
journal= {arXiv preprint arXiv:1704.06731},
year = {2018}
}