2017/11/06 by Alexandre Défossez, Francis Bach, Défossez, Alexandre +1 · 1 citation
Computer Science · Engineering · Mathematics · #Algorithm #Computer science #Convergence (economics) #FOS: Computer and information sciences #FOS: Mathematics #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Mathematical optimization #Mathematics #Operator (biology) #Optimization and Control (math.OC) #Parallel computing #Parallelizable manifold #Privacy-Preserving Technologies in Data #Regular polygon #Sparse and Compressive Sensing Techniques #Speedup #Stochastic Gradient Optimization Techniques #Variance (accounting) #cs.LG #math.OC #stat.ML
paper · pdf · doi:10.48550/arxiv.1711.01761
arxiv created 2017/11/06 · openalex publication_date 2017/11/06 · arxiv updated 2017/11/07 · openalex created_date 2022/09/01 · openalex updated_date 2026/08/05
We study a new aggregation operator for gradients coming from a mini-batch for stochastic gradient (SG) methods that allows a significant speed-up in the case of sparse optimization problems. We call this method AdaBatch and it only requires a few lines of code change compared to regular mini-batch SGD algorithms. We provide a theoretical insight to understand how this new class of algorithms is performing and show that it is equivalent to an implicit per-coordinate rescaling of the gradients, similarly to what Adagrad methods can do. In theory and in practice, this new aggregation allows to keep the same sample efficiency of SG methods while increasing the batch size. Experimentally, we also show that in the case of smooth convex optimization, our procedure can even obtain a better loss when increasing the batch size for a fixed number of samples. We then apply this new algorithm to obtain a parallelizable stochastic gradient method that is synchronous but allows speed-up on par with Hogwild! methods as convergence does not deteriorate with the increase of the batch size. The same approach can be used to make mini-batch provably efficient for variance-reduced SG methods such as SVRG.