Video Streaming with AI
AI & Media
The modern AI revolution (deep learning) has transformed the landscape of many computing domains, including computer networks and systems. Almost everybody is using video streaming apps today (TikTok, Youku, YouTube, Netflix, etc.). In this workshop we will learn all the behind-the-scenes technologies that go into building a scalable and high-quality streaming system, that is optimized with AI techniques. 

In this workshop the students will learn the technologies and components of a video streaming system that is built with the modern Dynamic Adaptive Streaming over HTTP (DASH) industry standard. Students will learn both in theory and in a practical, hands-on project about video encoding, video formats and representations, video streaming protocols, and video player optimzations. The workshop will also cover how video streaming systems use AI algorithms to make sure that each user has a great experience (called Quality of Experience, or QoE). The students will build their own AI-driven quality adaptation logic in Python. Student teams will then be able to compare their algorithm against others in a grand challenge competition and see who's AI system comes out on top.

By the end of this workshop, students will: 
  • Understand how modern video streaming systems work 
  • Understand the DASH streaming protocol that is used by most providers 
  • Understand how the quality adaptation logic works 
  • Build their own AI-driven quality adaptation logic as part of an end-to-end streaming system 
  • Work in teams, present their work and compare how well their AI-logic performs against the other teams
About Professor
Professor Roger Zimmermann
Department of Computer Science, School of Computing, NUS

Prof. Roger Zimmermann is a professor with the Computer Science Department at National University of Singapore (NUS). He is also a project lead with the Grab-NUS AI Lab and a deputy director with the Smart Systems Institute (SSI) at NUS. He holds a Ph.D. and an M.S. degree in Computer Science from the University of Southern California (USC).  

Among his research interests are streaming media and AR/VR architectures, dynamic adaptive streaming over HTTP (DASH), software defined networking (SDN), applications of machine learning, mobile location-based services, and spatio-temporal information management. Prof. Zimmermann has co-authored seven patents and more than three-hundred peer-reviewed articles in the above-mentioned research areas.

He is also involved in the multimedia community and was recently one of the directors of the IEEE Multimedia Communications Technical Committee (MMTC) Review Board. He is on the editorial board of the Springer International Journal of Multimedia Tools and Applications (MTAP) and an associate editor with the ACM Transactions on Multimedia journal (ACM TOMM) and IEEE MultiMedia. 

Prof. Zimmermann has been teaching a successful course on video streaming for more than a decade. He and his TA for this course, Dr. Abdelhak Bentaleb, have received a number of awards from the DASH Industry Forum for their work in video streaming (DASH-IF Excellence in DASH Awards 2018 and 2019).