Locality-aware and Fault-tolerant Batching for Machine Learning on Distributed Datasets | IEEE Journals & Magazine | IEEE Xplore