GaP: a factor model for discrete data

By John F. Canny. 2004

224 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Probabilistic & Bayesian Methods

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 Probabilistic & Bayesian Methods

Citations in the corpus

Rethinking LDA: Moment Matching for Discrete ICA (2015, neutral): “This model was introduced by Canny [12] and later named as a discrete ICA model [13].”
How algorithmic confounding in recommendation systems increases homogeneity and decreases utility (2017, background): “tion of non-negative matrix factorization [9] assumes that they are gamma-distributed.”
Automatic Differentiation Variational Inference (2016, methodology): “The Gamma Poisson factorization model describes discrete data matrices (Canny, 2004; Cemgil, 2009).”
Deep Exponential Families (2014, methodology): “We can represent a document as a vector of term counts modeled with Poisson random variables [9].”
Automatic Variational Inference in Stan (2015, methodology): “We continue by exploring two nonconjugate non-negative matrix factorization models: a constrained Gamma Poisson model [24] and a Dirichlet Exponential model.”
A Spectral Algorithm for Latent Dirichlet Allocation (2012, unknown): “Remark 7 (Skewed and Kurtotic ECA for LDA).”
Unsupervised Object Discovery: A Comparison (2010, methodology): “Latent Dirichlet Allocation (Blei et al. 2003) is a Bayesian version of the earlier Probabilistic Latent Semantic Analysis of Hofmann (1999), and (except when the priors are being optimized as well) both are versions of the general GammaPoisson model of Canny (2004).”
Learning author-topic models from text corpora (2010, methodology): “For example, Cohn and Hofmann [2001] proposed an extension of pLSI to model both the document content as well as citations or hyperlinks between documents. Similarly, Erosheva et al. [2004] extended the LDA model to model both text and citations and applied their model to scientific papers from the ”
Content-based recommendations with Poisson factorization (2014, methodology): “In both applications Poisson factorization has been shown to outperform competing methods [3, 9].”
Bayesian Nonparametric Poisson Factorization for Recommendation Systems (2014, methodology): “Finally, we remark that although we focus on recommendation systems, our BNP model can be used in a broader range of applications requiring matrix factorization, such as topic modeling (Canny, 2004) or community detection (Ball et al.”
Scalable Recommendation with Poisson Factorization (2013, methodology): “We model these data with factorized Poisson distributions [4], where each item i is represented by a vector of K latent attributes βi and each user u by a vector of K latent preferences βu.”
Probabilistic factor models for web site recommendation (2011, methodology): “GaP: this approach is proposed in [4], and it is originally used to model text mining problems.”
Beta-Negative Binomial Process and Poisson Factor Analysis (2011, methodology): “5 and shown in Table 1, NMF (Lee and Seung, 2000) is a special case of Γ-PFA; GaP (Canny, 2004) is a special case of β Γ-PFA; and in terms of inference, Dir-PFA is equivalent to LDA (Blei et al., 2003; Hoﬀman et al., 2010); and S γ Γ-PFA is closely related to FTM (Williamson et al., 2010).”
Scalable Recommendation with Hierarchical Poisson Factorization (2015, methodology): “Placing a Gamma prior on the user weights results in the GaP model [4], which was developed as an alternative text model to latent Dirichlet allocation (LDA) [3, 18].”
A Probabilistic Model for Using Social Networks in Personalized Item Recommendation (2015, methodology): “PF relates to the GaP topic model [5], and can be viewed as a type of Bayesian non-negative matrix factorization [21].”
Fast maximum margin matrix factorization for collaborative prediction (2005, methodology): “A common approach to collaborative prediction is to fit a factor model to the data, and use it in order to make further predictions (Azar et al., 2001; Billsus & Pazzani, 1998; Hofmann, 2004; Marlin & Zemel, 2004; Canny, 2004).”