Bayesian dual neural networks for recommendation

作     者：Jia HE Fuzhen ZHUANG Yanchi LIU Qing HE Fen LIN 

作者机构：Key Lab of Intelligent Information Processing of Chinese Academy of SciencesInstitute of Computing TechnologyChinese Academy of SciencesBeijing 100190China University of Chinese Academy of SciencesBeijing 100049China Rutgers UniversityNewark 07102USA Search Product CenterWeChat Search Application DepartmentTencentBeijing 100080China 

出 版 物：《Frontiers of Computer Science》 (中国计算机科学前沿（英文版）)

年 卷 期：2019年第13卷第6期

页      面：1255-1265页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：supported by the National Key R&D Program of China(2018YFB1004300) the National Natural Science Foundation of China(Grant Nos.61773361,61473273,91546122) the Science and Technology Project of Guangdong Province(2015B010109005) the Project of Youth Innovation Promotion Association CAS(2017146) supported by the funding of WeChat cooperation project.We thank Bo Che 

主　　题：collaborative filtering Bayesian neural network hybrid recommendation algorithm 

摘      要：Most traditional collaborative filtering(CF)methods only use the user-item rating matrix to make recommendations,which usually suffer from cold-start and sparsity *** address these problems,on the one hand,some CF methods are proposed to incorporate auxiliary information such as user/item profiles;on the other hand,deep neural networks,which have powerful ability in learning effective representations,have achieved great success in recommender ***,these neural network based recommendation methods rarely consider the uncertainty of weights in the network and only obtain point estimates of the ***,they maybe lack of calibrated probabilistic predictions and make overly confident *** this end,we propose a new Bayesian dual neural network framework,named BDNet,to incorporate auxiliary information for ***,we design two neural networks,one is to learn a common low dimensional space for users and items from the rating matrix,and another one is to project the attributes of users and items into another shared latent *** that,the outputs of these two neural networks are combined to produce the final ***,we introduce the uncertainty to all weights which are represented by probability distributions in our neural networks to make calibrated probabilistic *** experiments on real-world data sets are conducted to demonstrate the superiority of our model over various kinds of competitors.