Adaptive Video Streaming Based on Learning Intrinsic Reward

2022 IEEE International Symposium on Broadband Multimedia Systems and Broadcasting (BMSB)(2022)

引用 2|浏览3
暂无评分
摘要
The adaptive bitrate (ABR) algorithm based on reinforcement learning (RL) can actively learn bitrate control policies and adapt to different network environments. However, the user quality of experience (QoE) as the optimization objective contains multiple indicators. In the case of complex task objective, whether the reward feedback from the environment can effectively guide the policy update is the challenge of RL-based ABR algorithm. We propose an ABR algorithm based on learning intrinsic reward, which encourages exploration by enhancing the agent's intrinsic motivation and enables the agent to understand transitions between states to make an effective policy. The results show that our approach improves the average QoE value.
更多
查看译文
关键词
adaptive video streaming,reinforcement learning,intrinsic reward
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要