Improving the Accuracy-Latency Trade-off of Edge-Cloud Computation Offloading for Deep Learning Services | IEEE Conference Publication | IEEE Xplore