RecVAE: a New Variational Autoencoder for Top-N Recommendations with Implicit Feedback

TL;DR

RecVAE is a novel variational autoencoder model for top-N recommendation systems that introduces a composite prior, a new hyperparameter setting approach, and an alternating training method, significantly outperforming previous autoencoder models.

Contribution

RecVAE advances collaborative filtering by integrating a composite prior, a novel hyperparameter strategy, and an alternating training scheme, improving recommendation accuracy.

Findings

RecVAE outperforms Mult-VAE and RaCT on standard datasets.

Ablation studies confirm the effectiveness of each new component.

The model achieves state-of-the-art results in top-N recommendation tasks.

Abstract

Recent research has shown the advantages of using autoencoders based on deep neural networks for collaborative filtering. In particular, the recently proposed Mult-VAE model, which used the multinomial likelihood variational autoencoders, has shown excellent results for top-N recommendations. In this work, we propose the Recommender VAE (RecVAE) model that originates from our research on regularization techniques for variational autoencoders. RecVAE introduces several novel ideas to improve Mult-VAE, including a novel composite prior distribution for the latent codes, a new approach to setting the $\beta$ hyperparameter for the $\beta$-VAE framework, and a new approach to training based on alternating updates. In experimental evaluation, we show that RecVAE significantly outperforms previously proposed autoencoder-based models, including Mult-VAE and RaCT, across classical collaborative filtering datasets, and present a detailed ablation study to assess our new developments. Code and models are available at https://github.com/ilya-shenbin/RecVAE.