AdaEvo: Edge-Assisted Continuous and Timely DNN Model Evolution for Mobile Devices

Mobile video applications today have attracted sig nificant attention. Deep learning model (e.g., deep neural network, DNN) compression is widely used to enable on-device inference for facilitating robust and private mobile video applications. The compressed DNN, however, is vulnerable to the agnostic data drift of the live video captured from the dynamically changing mobile scenarios.Tocombatthedatadrift,mobileendsrelyonedgeservers to continuously evolve and re-compress the DNN with freshly collected data. We design a framework, AdaEvo, that efficiently supports the resource-limited edge server handling mobile DNN evolution tasks from multiplemobileends.ThekeygoalofAdaEvo is to maximizetheaveragequalityofexperience(QoE),i.e.,thepro portion of high-quality DNN service time to the entire life cycle, for all mobile ends. Specifically, it estimates the DNN accuracy drops at the mobile end without labels and performs a dedicated video frame sampling strategy to control the size of retraining data. In addition, it balances the limited computing and memory resources ontheedgeserverandthecompetitionbetweenasynchronoustasks initiated by different mobile users. With an extensive evaluation of real-world videos from mobile scenarios and across four diverse mobile tasks, experimental results show that AdaEvoenablesupto 34%accuracy improvement and 32% average QoE improvement.