基于深度学习推荐系统的研究与展望

Research and Prospect on Deep Learning Based Recommendation System

随着互联网信息不断呈指数增长,推荐系统成为了缓解信息过载的有效工具。深度学习作为目前的一个重要的研究方向,在许多研究领域都取得了突破性的进展。现有的研究表明,将深度学习技术融入到推荐过程中,可以通过整合海量的多源异构数据,使得用户模型更加贴合用户的偏好需求,从而提高推荐系统的性能和用户的满意度,并减轻信息过载的问题。文中对目前基于深度学习的推荐系统的相关研究进行全面的总结,首先阐述了传统推荐系统的内涵及其存在问题,然后详细介绍国内外学者通过深度学习解决上述问题的方法和策略,最后对深度学习在推荐领域的未来发展趋势进行分析和展望。

With the exponential growth of information on the Internet,the recommendation system has become an effective tool for alleviating the problem of information overload.As an important research direction at present,researches of deep learning have made breakthrough achievements in many fields.Meanwhile,recent studies also demonstrate its effectiveness in coping with information retrieval and recommendation system.It is a good way to apply deep learning techniques into recommendation systems,which can integrate the massive multi-source heterogeneous data to build more suitable user models according to user preferences requirements.The existing researches show that deep learning based recommendation system can improve the performance and user satisfaction and can alleviate the problem of information overload.In this paper,we have investigated the relevant literature of the current deep learning based recommendation systems.First,we elaborated the basic method of the traditional recommendation system and its problems.Then,we give an overview of the main deep learning techniques and introduce the application of deep learning techniques in the field of recommendation system by domestic and foreign scholars in details.Finally,we summarize the future development trend of deep learning based recommendation systems and make an outlook of the future work.