Toward Mobility-Aware Edge Inference Via Model Partition and Service Migration | IEEE Conference Publication | IEEE Xplore