Some extensions of score matching

By Aapo Hyvärinen. 2007

287 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Statistics and Probability, 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 Statistics and Probability

In Probabilistic & Bayesian Methods

Citations in the corpus

Neural Autoregressive Distribution Estimation (2016, background): “Other proposals include pseudo-likelihood (Besag, 1975; Marlin et al., 2010), score matching (Hyv¨arinen, 2005; Hyv¨arinen, 2007a,b), noise contrastive estimation (Gutmann and Hyv¨arinen, 2010) and probability ﬂow minimization (Sohl-Dickstein et al., 2011).”
Gradient-free Hamiltonian Monte Carlo with Efficient Kernel Exponential Families (2015, background): “This representation is of a form similar to [20, Section 4.1], the main differences being that the basis functions are chosen randomly, the basis set grows with n , and we will require an additional regularising term.”
Density Estimation in Infinite Dimensional Exponential Families (2013, methodology): “To counter the disadvantages of KDE and pseudo-MLE, in this paper, we propose to use the score matching method introduced by Hyv¨arinen [21, 22], which provides an eﬃcient way to address the problem at hand.”
Proper local scoring rules (2011, methodology): “Hyvärinen (2005) proposed a method for approximating a distribution P on X = R by a distribution Q in a specified family P of distributions by”
What regularized auto-encoders learn from the data-generating distribution (2012, methodology): “This notion has already been successfully applied to obtain the equivalent of score matching for discrete models, namely ratio matching (Hyvärinen, 2007).”
A Two-Layer Model of Natural Stimuli Estimated with Score Matching (2010, methodology): “Score matching (Hyvarinen, 2005 , 2007a , 2007b ) is an estimation method that allows learning statistical models that are specified only up to a multiplicative normalization constant (partition function).”
Clustering via Mode Seeking by Direct Estimation of the Gradient of a Log-Density (2014, background): “As in score matching for an exponential family [15], the optimization problem (2) can be solved analytically as”
How to Train Your Energy-Based Models (2021, background): “Score Matching assumes a continuous data distribution with positive density over the space, but it can be generalized to discrete or bounded data distributions (Hyvärinen, 2007; Lyu, 2012).”
Learning structured densities via infinite dimensional exponential families (2015, methodology): “To get around this problem, we propose to use score matching [10, 11] as a divergence, instead of the usual KL divergence, as it does not require evaluation of the log partition function.”
Regularized estimation of image statistics by Score Matching (2010, background): “Only some specific special-case solutions exist, e.g. for non-negative data [11].”
Learning Deep Architectures for AI (2007, background): “Another alternative to Contrastive Divergence is Score Matching [84, 85, 86], a general approach to train energy-based models in which the energy can be computed tractably, but not the normalization constant Z.”
Unsupervised Feature Learning and Deep Learning: A Review and New Perspectives (2012, background): “In particular, these include pseudo-likelihood (Besag, 1975) and ratiomatching (Hyvärinen, 2007).”
Structure Learning in Graphical Modeling (2016, background): “The score matching approach of Hyv¨arinen (2005, 2007) is well suited to address this challenge.”
PROPER LOCAL SCORING RULES ON DISCRETE SAMPLE SPACES (2011, methodology): “Replacing the log score with the Brier score leads to the method of ratio matching [Hyvärinen (2007)].”