Dual-embedding based Neural Collaborative Filtering for Recommender Systems

TL;DR

This paper introduces DNCF, a dual-embedding neural collaborative filtering framework that enhances user and item representations by leveraging historical interactions, leading to improved recommendation accuracy.

Contribution

The paper proposes a novel dual-embedding approach for collaborative filtering that incorporates interaction history to improve embedding quality.

Findings

DNCF outperforms traditional matrix factorization models.

DNCF surpasses other state-of-the-art deep learning recommender models.

Extensive experiments validate the effectiveness of dual-embedding in CF.

Abstract

Among various recommender techniques, collaborative filtering (CF) is the most successful one. And a key problem in CF is how to represent users and items. Previous works usually represent a user (an item) as a vector of latent factors (aka. \textit{embedding}) and then model the interactions between users and items based on the representations. Despite its effectiveness, we argue that it's insufficient to yield satisfactory embeddings for collaborative filtering. Inspired by the idea of SVD++ that represents users based on themselves and their interacted items, we propose a general collaborative filtering framework named DNCF, short for Dual-embedding based Neural Collaborative Filtering, to utilize historical interactions to enhance the representation. In addition to learning the primitive embedding for a user (an item), we introduce an additional embedding from the perspective of the interacted items (users) to augment the user (item) representation. Extensive experiments on four publicly datasets demonstrated the effectiveness of our proposed DNCF framework by comparing its performance with several traditional matrix factorization models and other state-of-the-art deep learning based recommender models.