Abstract:Social recommender systems are expected to improve recommendation quality by incorporating social information when there is little user-item interaction data. However, recent reports from industry show that social recommender systems consistently fail in practice. According to the negative findings, the failure is attributed to: (1) A majority of users only have a very limited number of neighbors in social networks and can hardly benefit from social relations; (2) Social relations are noisy but they are indiscriminately used; (3) Social relations are assumed to be universally applicable to multiple scenarios while they are actually multi-faceted and show heterogeneous strengths in different scenarios. Most existing social recommendation models only consider the homophily in social networks and neglect these drawbacks. In this paper we propose a deep adversarial framework based on graph convolutional networks (GCN) to address these problems. Concretely, for (1) and (2), a GCN-based autoencoder is developed to augment the relation data by encoding high-order and complex connectivity patterns, and meanwhile is optimized subject to the constraint of reconstructing the social profile to guarantee the validity of the identified neighborhood. After obtaining enough purified social relations for each user, a GCN-based attentive social recommendation module is designed to address (3) by capturing the heterogeneous strengths of social relations. Finally, we adopt adversarial training to unify all the components by playing a Minimax game and ensure a coordinated effort to enhance recommendation performance. Extensive experiments on multiple open datasets demonstrate the superiority of our framework and the ablation study confirms the importance and effectiveness of each component.

A privacy preserving recommendation and fraud detection method based on graph convolution

Decentralized Graph Neural Network for Privacy-Preserving Recommendation

Graph Embedding for Recommendation against Attribute Inference Attacks

A GCN-based Model for Recommendation Using Local Differential Privacy Method

FedGNN: Federated Graph Neural Network for Privacy-Preserving Recommendation

IntentGC: a Scalable Graph Convolution Framework Fusing Heterogeneous Information for Recommendation

Lossless and Privacy-Preserving Graph Convolution Network for Federated Item Recommendation

Towards Adversarially Robust Recommendation from Adaptive Fraudster Detection

Enhancing Social Recommendation with Adversarial Graph Convolutional Networks

Subgraph-level federated graph neural network for privacy-preserving recommendation with meta-learning

FedRKG: A Privacy-preserving Federated Recommendation Framework via Knowledge Graph Enhancement

Privacy-Preserving Graph Convolutional Networks for Text Classification

Robust Graph Embedding Recommendation Against Data Poisoning Attack.

Towards Personalized Privacy: User-Governed Data Contribution for Federated Recommendation

Graph-Refined Convolutional Network for Multimedia Recommendation with Implicit Feedback

Federated Heterogeneous Graph Neural Network for Privacy-preserving Recommendation

Cluster-Enhanced Federated Graph Neural Network for Recommendation

How Powerful is Graph Convolution for Recommendation?

Vertical Federated Graph Neural Network for Recommender System

Data Privacy Protection Model Based on Graph Convolutional Neural Network

Dynamically Expandable Graph Convolution for Streaming Recommendation