电子学报 ›› 2020, Vol. 48 ›› Issue (9): 1672-1679.DOI: 10.3969/j.issn.0372-2112.2020.09.002

• 学术论文 • 上一篇    下一篇

基于多层注意力表示的音乐推荐模型

李琳, 唐守廉   

  1. 北京邮电大学经济管理学院, 北京 100876
  • 收稿日期:2020-02-05 修回日期:2020-04-09 出版日期:2020-09-25
    • 通讯作者:
    • 李琳
    • 作者简介:
    • 唐守廉 男,1952年出生于上海,北京邮电大学经济管理学院教授、博士生导师,主要研究为政府电信管制政策研究,企业发展战略研究,企业市场营销策划.E-mail:tangshoulian@263.net

Hierarchical Attention Representation Model for Music Recommendation

LI Lin, TANG Shou-lian   

  1. School of Economics and Management, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • Received:2020-02-05 Revised:2020-04-09 Online:2020-09-25 Published:2020-09-25
    • Corresponding author:
    • LI Lin

摘要: 目前的音乐推荐方法只能挖掘用户与歌曲之间的一般性关系,无法区分不同用户对同一首歌曲的差异性偏好.为此,本文提出了基于多层注意力表示的音乐推荐模型,利用用户属性信息和歌曲内容信息从多维度学习歌曲表征,挖掘用户与歌曲之间的偏好关系.为了区分用户对歌曲多域特征的差异性偏好,设计了用户特征依赖的注意力网络;为了区分不同历史行为对用户偏好的差异性,挖掘用户行为的时序依赖关系,设计了歌曲依赖的注意力网络.最后,利用Softmax函数计算用户对候选歌曲的偏好分布并产生推荐.在30Music和MIGU数据集上的实验结果表明,相比目前的推荐模型,本文提出的模型在Recall和MRR均得到了显著提升.

关键词: 特征表示, 注意力网络, 时序关系, 音乐推荐

Abstract: Current music recommendation models can mine the general preference between users and music, which are unable to distinguish the differential preference of different users towards the same song. Therefore, we propose a hierarchical attention representation model (HARM) to improve the music recommendation quality. HARM utilizes attributes of users and contents of music to learn music representation from the perspective of multi-dimension and mine the preference relationship between user and music. In order to mine users' differential preference on music's multi-field feature, a user feature-dependent attention network is designed. In addition, in order to mine the impacts of different historical behavior on user preference and learn sequential dependency of user's, a music-dependent attention network is designed. Finally, recommendation is generated by using a softmax function to calculate the preference distribution of users on candidate songs. The experimental results on 30Music and MIGU datasets shows that, comparing with the existent recommendation models, HARM can gain significant improvement on Recall and MRR.

Key words: feature representation, attention network, sequential relationship, music recommendation

中图分类号: