关于人工智能对用户QoE提升的若干思考

关于人工智能对用户QoE提升的若干思考

高赟，魏昕，周亮

南京邮电大学

摘要：认为网络视频供应商也开始更加关注用户体验质量（QoE），但仍存在着数据量大、个性化提升用户QoE困难等问题。梳理了人工智能（AI）提升大数据时代下用户QoE的系统化解决方案。以视频业务中的流媒体码率自适应任务为例，将基于用户行为分析的个性化QoE模型及时“反馈”到深度学习神经网络中，通过模型自适应决策和码率调节，实现用户QoE最大化。最后指出基于人工智能的QoE提升仍然存在一些挑战。
关键词：QoE；AI；多媒体通信

QoE Improvement Issues Based on Artificial Intelligence

GAO Yun, WEI Xin, ZHOU Liang

Nanjing University of Posts and Telecommunications

Abstract: Online video providers are more concerned about users’ Quality of Experience (QoE). However, there are still some problems in multimedia domain such as large amount of user data and user personalized experience improvement. A systematic solution for improving user’ QoE by Artificial Intelligence (AI) technology in the era of big data is provided. In this paper, the adaptive streaming media rate of video services is taken as an example. In order to maximize the users’ QoE, the personalized QoE model based on user behavior analysis timely provide feedback into the deep learning neural network, and then adjust media rate by making the model adaptively decision-making. Finally, we point out the challenges of QoE improvement under AI technology.
Keywords: QoE, AI, multimedia communication