JointRec: A Deep-Learning-Based Joint Cloud Video Recommendation Framework for Mobile IoT

Abstract

In the era of Internet of Things (IoT), watching videos on mobile devices has been a popular application in our daily life. How to recommend videos to users is one of the most concerned problem for Internet video service providers (IVSPs). In order to provide better recommendation service to users, they deploy cloud servers in a geo-distributed manner. Each server is responsible for analyzing a local area of user data. Therefore, these cloud servers form information islands and the characteristics of data present nonindependent and identically distribution (non-i.i.d). In this scenario, it is difficult to provide accurate video recommendation service to the minority of users in each area. To tackle this issue, we propose JointRec, a deep learning-based joint cloud video recommendation framework. JointRec integrates the JointCloud architecture into mobile IoT and achieves federated training among distributed cloud servers. Specifically, we first design a dual-convolutional probabilistic matrix factorization (Dual-CPMF) model to conduct video recommendation. Based on this model, each cloud can recommend videos by exploiting the user's profiles and description of videos that users rate, thereby providing more accurate video recommendation services. Then, we present a federated recommendation algorithm which enables each cloud to share their weights and train a model cooperatively. Furthermore, considering the heavy communication costs in the process of federated training, we combine low-rank matrix factorization and 8-bit quantization method to reduce uplink communication costs and network bandwidth. We validate the proposed approach on the real-world data set, and the experimental results indicate the effectiveness of our proposed approach.