Fast Sparse Gaussian Process Methods: The Informative Vector Machine

By Neil D. Lawrence, Matthias W. Seeger, Ralf Herbrich. 2002

617 citations recorded by the source corpus.

How researchers see this paper

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

Citations in the corpus

Functional Regularisation for Continual Learning using Gaussian Processes (2019, methodology): “by memorising an approximate posterior belief over the underlying task-speciﬁc function. To implement this, we consider Gaussian processes (GPs) [23], and make use of inducing point sparse GP methods [6, 16, 33, 22, 35, 36, 12, 4], which summarise posterior distributions over functions using small n”
Scalable Variational Gaussian Process Classification (2014, methodology): “Related work The informative vector machine (IVM) [Lawrence et al., 2003] is the first work to approach sparse GP classification to our knowledge.”
Collaborative Gaussian Processes for Preference Learning (2012, methodology): “In practice, current approaches make approximations for the computation of the posterior entropy [12, 9].”
Sparse Gaussian Processes using Pseudo-inputs (2005, background): “In recent years there have been many attempts to make sparse approximations to the full GP in order to bring this scaling down to MN whereM N [1, 2, 3, 4, 5, 6, 7, 8, 9].”
Variable Noise and Dimensionality Reduction for Sparse Gaussian processes (2006, methodology): “een many methods proposed in recent years to address this problem, and bring the scaling down to O(M2N) where M ˝ N [Tresp, 2000, Smola and Bartlett, 2001, Williams and Seeger, 2001, Csat´o, 2002a,b, Lawrence et al., 2003, Seeger et al., 2003, Seeger, 2003, Quin˜onero Candela, 2004]. Recently we dev”
Gaussian Processes for Prediction (2007, methodology): “[2005] use the informative vector machine [Lawrence et al., 2003] to select a subset of training points that is deemed most informative.”
A tutorial on Gaussian process regression: Modelling, exploring, and exploiting functions (2016, methodology): “Some methods that have been proposed are to sparsely approximate inputs (Lawrence, Seeger, and Herbrich, 2003) or to bound the computational cost of the matrix inversion by projecting into a pre-defined finite basis of functions drawn from the eigen-spectrum of the kernel (Rahimi and Recht, 2007).”
Scalable Inference for Gaussian Process Models with Black-Box Likelihoods (2015, background): “There has been a long-standing interest in the GP community to overcome the cubic scaling of inference in standard GP models [17, 18, 12, 13, 8].”
An unbiased offline evaluation of contextual bandit algorithms with generalized linear models (2011, methodology): “To approximate the posterior distributions, we carry out the Laplace approximation for logistic models, and the assumed density filtering or expectation propagation (Lawrence et al., 2002; Minka, 2001) for probit.”
Few-Shot Self Reminder to Overcome Catastrophic Forgetting (2018, background): “Finally for point selection, although both our and other attempts in the literature find class-stratified sampling to very effective, it would be useful to further investigate the parallels to GP sparsification techniques [4, 11], where the goal is also to select points which provide significant inf”
Sparse Convolved Gaussian Processes for Multi-output Regression (2008, methodology): “For computational speed up the informative vector machine (IVM) is employed [8].”
Computationally Efficient Convolved Multiple Output Gaussian Processes (2011, methodology): “For computational speed up the informative vector machine (IVM) is employed (Lawrence et al., 2003).”