Professor Shaolei Ren received a National Science Foundation grant to study automated design of deep neural networks (DNNs) for edge inference. Edge devices, such as mobile phones, drones and robots, have been emerging as an increasingly more important platform for DNN inference. But, designing an optimal DNN model for maximizing the users' quality of experience (QoE) is significantly challenged by the high degree of heterogeneity in edge devices and constantly-changing usage scenarios. Moreover, the existing approaches focus on optimizing a certain objective metric for edge inference, which may not translate into improvement of the actual QoE for users. Keeping users into a closed loop, Professor Ren will develop novel online learning techniques to automate the design of DNNs for QoE-optimal edge inference.

This is a collaborative project between UCR and the University of Miami, and UCR is the lead institution.