Latent semantic models for collaborative filtering

By Thomas Hofmann. 2004

1,513 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Information Systems

How researchers see this paper

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

In Artificial Intelligence

In Information Systems

Citations in the corpus

Recurrent Latent Variable Networks for Session-Based Recommendation (2017, background): “Memory-based systems generate predictions under a neighborhood-driven rationale: Item rating prediction for a target user comprises determination of other users with similar ratings (target user neighbors), and computation of a weighted average of the ratings of each item provided by the target user”
BPR: Bayesian Personalized Ranking from Implicit Feedback (2009, methodology): “MF models are known to outperform [12] many other models including the Bayesian models URP [9] and PLSA [4] for the related task of collaborative rating prediction.”
SoRec: social recommendation using probabilistic matrix factorization (2008, methodology): “Examples of model-based approaches include the clustering model [25], aspect models [8, 9, 20] and the latent factor model [3].”
Learning to recommend with social trust ensemble (2009, methodology): “Examples of model-based approaches include aspect models [7, 8, 21], the latent factor model [4], the Bayesian hierarchical model [24] and the ranking model [11].”
Recommender Systems (2016, positive): “The early works [133, 252, 300, 500, 569, 666] showed how different forms of matrix factorization could be used for recommendations.”
A generalized maximum entropy approach to bregman co-clustering and matrix approximation (2004, methodology): “14 shows the mean absolute error (MAE) obtained using various existing collaborative filtering approaches (Sarwar et al., 2000; Hofmann, 2004; Resnick et al., 1994) as well as the co-clustering approach based on squared Euclidean distance.”
Nonparametric bayesian multitask collaborative filtering (2013, methodology): “This is the earliest kind of algorithms used in the context of CF systems, and still comprises the basis of most of the filtering functionality performed on popular websites such as Amazon [10, 1].”
Learning from Incomplete Ratings Using Non-negative Matrix Factorization (2006, background): “Generally, they can be divided into two categories: memory based algorithms [1, 6, 16, 17] and model based algorithms [2, 3, 5, 7, 11, 15, 18, 19, 20].”
A non-IID Framework for Collaborative Filtering with Restricted Boltzmann Machines (2013, methodology): “Related model-based approaches include statistical latent models (Hofmann, 2004), uncertain graphs (Taranto et al., 2012) and autoassociative networks (Vozalis et al., 2010).”
An automatic weighting scheme for collaborative filtering (2004, methodology): “In [14], the ratings of each user are normalized to be norm distribution with zero mean and one variance.”
Improving maximum margin matrix factorization (2008, methodology): “For example by extracting a feature vector for each user and item in the data set such that the inner product of these features minimizes an explicit or implicit loss functional (e.g. Hoffman 2004 following a probabilistic approach).”
An Efficient Non-Negative Matrix-Factorization-Based Approach to Collaborative Filtering for Recommender Systems (2014, methodology): “Hofmann employs probabilistic latent semantic analysis to build a CF model [10]; Srebro et al. propose the maximum-margin MF-based model [11]; Kurucz et al. propose the expectation maximization (EM)based MF model [12].”
Advances in Collaborative Filtering (2011, background): “Latent factor models approach collaborative filtering with the holistic goal to uncover latent features that explain observed ratings; examp les include pLSA [15],”
Collaborative prediction using ensembles of Maximum Margin Matrix Factorizations (2006, methodology): “A linear factor model is a simple, natural, and well-studied (e.g. (Billus & Pazzani, 1998), (Hofmann, 2004),(Marlin & Zemel, 2004)) approach to collaborative prediction: (cid:12)t the n -by- d ratings matrix by a low rank decomposition UV 0 , where U is n -by- k and V is d -by- k and k min( n; d ).”