Optimizing bulk synchronous parallel ML | IEEE Conference Publication | IEEE Xplore