DGCN: Diversified Recommendation with Graph Convolutional Networks

TL;DR

This paper introduces DGCN, a novel recommendation approach that integrates diversification into the candidate generation process using Graph Convolutional Networks, enhancing both diversity and accuracy.

Contribution

It proposes a unified framework combining GCN with rebalanced neighbor discovering and adversarial learning to improve recommendation diversity at the candidate generation stage.

Findings

Significantly improves recommendation diversity.

Alleviates the accuracy-diversity trade-off.

Outperforms baseline methods on real-world datasets.

Abstract

These years much effort has been devoted to improving the accuracy or relevance of the recommendation system. Diversity, a crucial factor which measures the dissimilarity among the recommended items, received rather little scrutiny. Directly related to user satisfaction, diversification is usually taken into consideration after generating the candidate items. However, this decoupled design of diversification and candidate generation makes the whole system suboptimal. In this paper, we aim at pushing the diversification to the upstream candidate generation stage, with the help of Graph Convolutional Networks (GCN). Although GCN based recommendation algorithms have shown great power in modeling complex collaborative filtering effect to improve the accuracy of recommendation, how diversity changes is ignored in those advanced works. We propose to perform rebalanced neighbor discovering, category-boosted negative sampling and adversarial learning on top of GCN. We conduct extensive experiments on real-world datasets. Experimental results verify the effectiveness of our proposed method on diversification. Further ablation studies validate that our proposed method significantly alleviates the accuracy-diversity dilemma.