报告题目：基于机器学习的码率自适应算法再思考

报 告 人：王朔遥

报告时间：2024年4月18日 14:30-15:00

报告地点：计算机楼A505

报告摘要：

With the development of Ultra-High-Definition (UHD) video, Virtual Reality (VR), and other video streaming applications, video traffic is skyrocketing. Adaptive video streaming plays a crucial role in ensuring high-quality video streaming services. In this talk, we discuss how to improve the adaptive video streaming performance through the integration of machine learning and mechanism designs. In particular, we first discuss integrating pipeline mechanism design and reinforcement learning techniques to enable adaptive video streaming in multi-tier computing networks. Subsequently, we explore integrating stochastic model predictive control and time-series data prediction techniques to enhance user experience in 360-degree video streaming.

报告人简介：

王朔遥，2013年和2018年分别获香港中文大学信息工程学士学位和博士学位，现为深圳大学电子与信息工程学院副教授。当前主要研究方向为多媒体通信和边缘智能，担任通信领域国际权威期刊IEEE Transactions on Mobile Computing和IEEE Open Journal of Communications编委。