Abstract:Collaborative filtering is an effective and widely used recommendation approach by applying the user-item rating matrix for recommendations, however, which usually suffers from cold-start and sparsity problems. To address these problems, hybrid methods are proposed to incorporate auxiliary information such as user/item profiles to collaborative filtering models; Cross-domain recommendation systems add a new dimension to solve these problems by leveraging ratings from other domains to improve recommendation performance. Among these methods, deep neural network based recommendation systems achieve excellent performance due to their excellent ability in learning powerful representations. However, these cross-domain recommendation systems based on deep neural network rarely consider the uncertainty of weights. Therefore, they maybe lack of calibrated probabilistic predictions and make overly confident decisions. Along this line, we propose a general cross-domain recommendation framework via Bayesian neural network to incorporate auxiliary information, which takes advantage of both the hybrid recommendation methods and the cross-domain recommendation systems. Specifically, our framework consists of two kinds of neural networks, one to learn the low dimensional representation from the one-hot codings of users/items, while the other one is to project the auxiliary information of users/items into another latent space. The final rating is produced by integrating the latent representations of the one-hot codings of users/items and the auxiliary information of users/items. The latent representations of users learnt from ratings and auxiliary information are shared across different domains for knowledge transfer. Moreover, we capture the uncertainty in all weights by representing weights with Gaussian distributions to make calibrated probabilistic predictions. We have done extensive experiments on real-world data sets to verify the effectiveness of our framework.

Improving Cross-Domain Recommendation Through Probabilistic Cluster-Level Latent Factor Model

Cross-Domain Recommendation via Cluster-Level Latent Factor Model.

Improving Recommendations With Collaborative Factors

User Distribution Mapping Modelling with Collaborative Filtering for Cross Domain Recommendation

Cross-Domain Recommendation Algorithm Based on Latent Factor Model

Cross-Domain Recommendation Via User-Clustering and Multidimensional Information Fusion

A Multifaceted Model for Cross Domain Recommendation Systems.

Enhancing Cross-domain Recommendation Through Preference Structure Information Sharing

Deep Low-rank Sparse Collective Factorization for Cross-Domain Recommendation

Cross-Domain Recommendation Algorithm Based on Individual Information

Pairwise cross-domain factor model for heterogeneous transfer ranking.

A Cross-Domain Recommendation Mechanism for Cold-Start Users Based on Partial Least Squares Regression

Federated Probabilistic Preference Distribution Modelling with Compactness Co-Clustering for Privacy-Preserving Multi-Domain Recommendation

Heterogeneous and Clustering-Enhanced Personalized Preference Transfer for Cross-Domain Recommendation

LSCD: Low-rank and Sparse Cross-Domain Recommendation

Personalized Recommendation Via Cross-Domain Triadic Factorization.

Cross-domain Recommendation by Combining Feature Tags with Transfer Learning

Cross-domain Recommendation Based on Latent Factor Alignment

CRTL: Context Restoration Transfer Learning for Cross-Domain Recommendations

A General Cross-Domain Recommendation Framework Via Bayesian Neural Network.

A cross-domain collaborative filtering algorithm with expanding user and item features via the latent factor space of auxiliary domains