Variational Autoencoders for Collaborative Filtering

With Mult-VAE, the authors introduce a generative model with multinomial likelihood, propose a different regularization parameter for the learning objective, and use Bayesian inference for parameter estimation.

B) KL Regularization Term Meaning

This means that we are less able to generalize to novel user clicks from historical data.

C) Improvement

Most works that develop further developments of VAE for collaborative filtering introduce alternative loss functions

D) Think

• the choice of the used cutofs (20 and 50 for Recall, and 100 for NDCG) is not very consistent in the paper.

E) Related

F) References

• Autoencoders for Collaborative Filtering - WSDM 2020 paper “RecVAE: A New Variational Autoencoder for Top-N Recommendations with Implicit Feedback”