Offloaded Execution of Deep Learning Inference at Edge: Challenges and Insights | IEEE Conference Publication | IEEE Xplore