MusicRoBot: Towards Conversational Context-Aware Music Recommender System
- 2.6k Downloads
Traditional recommendation approaches work well on depicting users’ long-term music preference. However, in the conversational applications, it is unable to capture users’ real time music taste, which are dynamic and depend on user context including users’ emotion, current activities or sites. To meet users’ real time music preferences, we have developed a conversational music recommender system based on music knowledge graph, MusicRoBot (Music RecOmmendation Bot). We embed the music recommendation into a chatbot, integrating both the advantages of dialogue system and recommender system. In our system, conversational interaction helps capture more real-time and richer requirements. Users can receive real time recommendation and give feedbacks by conversation. Besides, MusicRoBot also provides the music Q&A function to answer several types of musical question by the music knowledge graph. A WeChat based service has been deployed piloted for volunteers already.
KeywordsMusic recommendation Online recommendation Dialogue system Recommender system
- 1.Christakopoulou, K., Radlinski, F., Hofmann, K.: Towards conversational recommender systems. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 815–824. ACM (2016)Google Scholar
- 2.Sun, Y., Zhang, Y., Chen, Y., et al.: Conversational recommendation system with unsupervised learning. In: Proceedings of the 10th ACM Conference on Recommender Systems, pp. 397–398. ACM (2016)Google Scholar
- 3.Qin, L., Chen, S., Zhu, X.: Contextual combinatorial bandit and its application on diversified online recommendation. In: Proceedings of the 2014 SIAM International Conference on Data Mining, pp. 461–469. Society for Industrial and Applied Mathematics (2014)Google Scholar