Abstract: As Deep Neural Networks (DNNs) have been widely used in various applications, including computer vision on image segmentation and recognition, it is important to reduce the makespan of DNN inference computation, especially when running on mobile devices. Offloading is a viable solution that offloads computation from a slow mobile device to a fast, but remote edge/cloud. As DNN computation consists of a multiple-stage processing pipeline, it is critical to decide on what stage should offloading occur to minimize the makespan. Our observations show that the local computation time on a mobile device follows a linear increasing function, while the offloading time on a mobile device is monotonic decreasing and follows a convex curve as more DNN layers are computed in the mobile device. Based on this observation, we first study the optimal partition and scheduling for one line-structure DNN. Then, we extend the result to multiple line-structure DNNs. Heuristic results for general-structure DNNs, represented by Directed Acyclic Graphs (DAGs), are also elaborated based on a path-based scheduling policy. Extensions to DNN training are also discussed.

Short Bio: Jie Wu is the Director of the Center for Networked Computing and Laura H. Carnell professor at Temple University. He also serves as the Director of International Affairs at College of Science and Technology. He served as Chair of Department of Computer and Information Sciences from the summer of 2009 to the summer of 2016 and Associate Vice Provost for International Affairs from the fall of 2015 to the summer of 2017. Prior to joining Temple University, he was a program director at the National Science Foundation and was a distinguished professor at Florida Atlantic University. His current research interests include mobile computing and wireless networks, routing protocols, network trust and security, distributed algorithms, applied machine learning, and cloud computing. Dr. Wu regularly publishes in scholarly journals, conference proceedings, and books. He serves on several editorial boards, including IEEE Transactions on Mobile Computing, IEEE Transactions on Service Computing, Journal of Parallel and Distributed Computing, and Journal of Computer Science and Technology. Dr. Wu is/was general chair/co-chair for IEEE IPDPS’08, IEEE DCOSS’09, IEEE ICDCS’13, ACM MobiHoc’14, ICPP’16, IEEE CNS’16, WiOpt’21, and ICDCN’22 as well as program chair/cochair for IEEE MASS’04, IEEE INFOCOM’11, CCF CNCC’13, and ICCCN’20. He was an IEEE Computer Society Distinguished Visitor, ACM Distinguished Speaker, and chair for the IEEE Technical Committee on Distributed Processing (TCDP). Dr. Wu is a Fellow of the AAAS and a Fellow of the IEEE.