On the Distribution of ML Workloads to the Network Edge and Beyond | IEEE Conference Publication | IEEE Xplore