Scalable Recommendation with Hierarchical Poisson Factorization

By Prem Gopalan, Jake M. Hofman, David M. Blei. 2015

289 citations recorded by the source corpus.

Subfields: Probabilistic & Bayesian Methods, Recommender Systems

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Probabilistic & Bayesian Methods

In Recommender Systems

Citations in the corpus

Variational Autoencoders for Collaborative Filtering (2018, methodology): “For most of the latent factor collaborative filtering model, e.g., matrix factorization [13, 19], when given the click history of a user that is not present in the training data, normally we need to perform some form of optimization to obtain the latent factor for this user.”
A Probabilistic Model for Using Social Networks in Personalized Item Recommendation (2015, methodology): “(SPF is based on Poisson factorization [11], a new model that is particularly suited for implicit data.”
The Impact of Machine Learning on Economics (2018, methodology): “Not all dimensionality reduction methods involve creating clusters; older methods such as principal components analysis can be used to reduce dimensionality, while modern methods include matrix factorization (finding two low-dimensional matrices whose product well approximates a larger matrix), regu”
The Blessings of Multiple Causes (2018, background): “When the causes are counts, Poisson factorization (pf) is an appropriate factor model (Cemgil 2009; Schmidt, Winther, and Hansen 2009; Gopalan, Hofman, and Blei 2015). pf is a probabilistic form of nonnegative matrix factorization (Lee and Seung 1999, 2001), where z i and θ j are positive K…”
Session-Based Social Recommendation via Dynamic Graph Attention Networks (2019, methodology): “arned temporal representations by factorizing the (user, item, time) tensor. Koren [19] developed a similar model named timeSVD++. Charlin et al. [5] modeled similarly but using Poisson factorization [10]. However,theseapproachesassumethattheinterestofuserschanges slowlyandsmoothlyoverlong-termhoriz”
Exponential Family Embeddings (2016, methodology): “Mirror-ing the success of word embeddings, ef-emb models outperform traditional dimension reduction, such as exponential family principal component analysis ( pca ) (Collins et al., 2001) and Poisson factorization (Gopalan et al., 2015), and ﬁnd interpretable features of the data.”
Neural Survival Recommender (2017, methodology): “We compare our model with several strong baseline meth-ods, as described below: Hierarchical Poisson Factorization (HPF) [11] improves upon the traditional matrix factorization model by modeling the observed data as Poisson distributed instead of standard Gaussian variable.”