[Seminar] How will Deep Learning Change Internet Video Delivery?

Tuesday, November 27th 2018, 11:00am - Tuesday, November 27th 2018, 12:00pm
302동 107호

호스트: 전병곤 교수(x1928, 880-1928)


Internet video has experienced tremendous growth over the last few decades and is still growing at a rapid pace. Internet video now accounts for 73% of Internet traffic and is expected to quadruple in the next five years. Augmented reality and virtual reality streaming, projected to increase twentyfold in five years, will also accelerate this trend. In this talk, I will argue that advancement in deep neural networks presents new opportunities that can fundamentally change Internet video delivery. In particular, deep neural networks allow the content delivery network to easily capture the content of the video and thus enable content-aware video delivery. To demonstrate this, I will present NAS, a new Internet video delivery framework that integrates deep neural network based quality enhancements with adaptive streaming. NAS incorporates a super-resolution deep neural network (DNN) and a deep re-inforcement neural network to optimize the user quality of experience (QoE). It outperforms the current state of the art, enhancing the average QoE by 43.08% using the same bandwidth budget or saving 17.13% of bandwidth while providing the same user QoE. Finally, I will talk about future research ideas and new design space that the integration with deep learning and video streaming technology enables.

Speaker Bio

Bio: Dongsu Han is an associate professor in the School of Electrical Engineering at KAIST. He has actively worked in the area of systems and networking focusing on problems that arise from the fact that modern networking applications often run on the cloud at scale, such as high-speed network and application design, low-latency congestion control, and security and privacy of network applications. He has published numerous technical papers at premier venues, including SIGCOMM, OSDI, NSDI, CCS, Mobisys, CoNEXT, and EuroSys. Notable recognitions of his work include USENIX NSDI Best Paper Award and USENIX NSDI Community Award. He has served as a program committee member for a number of outstanding venues, including SIGCOMM, NSDI, HotNets, CoNEXT, and INFOCOM.