Sparse Online Gaussian Processes

By Manfred Opper. 2008

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

Information-theoretic Online Memory Selection for Continual Learning (2022, neutral): “Sparse Gaussian processes also consider the selection of inducing points (Seeger et al., 2003; Csato & Opper, 2002; Bui et al., 2017).”
Functional Regularisation for Continual Learning using Gaussian Processes (2019, methodology): “Within the online sparse Gaussian process approximation literature, the recent work of (Bui et al., 2017a) extends the two earlier seminal works of (Csato & Opper, 2002) and (Csato, 2002), unifying them within a power expectation framework and proposing methods for online learning of…”
Streaming Sparse Gaussian Process Approximations (2017, methodology): “For the case α = 1, which is called Expectation Propagation, we recover the seminal work by Csató and Opper [8].”
Partitioned Variational Inference: A unified framework encompassing federated and continual learning (2018, background): “…on pseudo-point sparse approximations [Titsias, 2009, Hensman et al., 2015, Matthews et al., 2016, Bui et al., 2017b] and streaming approximations [Csató and Opper, 2002, Bui et al., 2017a], we develop a streaming variational approximation that approximates the posterior distribution over both…”
Model learning for robot control: a survey (2011, methodology): “Sparse Gaussian process regression (Csato and Opper 2002) Global Incremental Yes Oðn(2)Þ Transition dynamics (Rottmann and Burgard 2009), Task model (Grollman and Jenkins 2008)”
Variational Bayesian Multinomial Probit Regression with Gaussian Process Priors (2006, background): “…a large body of literature focusing on reducing this cost via reduced rank approximations (Williams & Seeger, 2001) and sparse online learning (Csato & Opper, 2002; Quinonero-Candela & Winther, 2003) where assumed density ﬁltering(ADF)formsthebasisofonlinelearningandsparseapproximations…”
A Unifying Framework for Sparse Gaussian Process Approximation using Power Expectation Propagation (2016, methodology): “Astonishingly Csató and Opper not only developed such a method in 2002 [10], predating much of the work mentioned above, they showed that it is equivalent to applying the FITC approximation and running EP if further approximation is required.”
Online Sparse Multi-Output Gaussian Process Regression and Learning (2019, methodology): “In [47], a technique called sparse online GP (SOGP) was proposed for Fig.”
Local and global sparse Gaussian process approximations (2007, background): “In recent years there have been many sparse GP approximations developed that reduce the cost to O(N M 2 ) training time and O(M 2 ) prediction time [e.g. Csató and Opper, 2002, Seeger et al., 2003, Snelson and Ghahramani, 2006].”
Approximation Methods for Gaussian Process Regression (2007, methodology): “Csató and Opper (2002) present and " online " method, where the examples are processed sequentially, while Csató et al. (2003) give an EP type algorithm where multiple sweeps across the data are permitted.”
Variational Learning of Inducing Variables in Sparse Gaussian Processes (2009, methodology): “Clearly, the predictive distribution is exactly the one used by the projected process (PP) that has been previously proposed in (Csato and Opper, 2002; Seeger et al., 2003).”
Efficient Sampling for Gaussian Process Inference using Control Variables (2008, methodology): “In regression problems with Gaussian likelihoods, inference in GP models is analytically tractable, while for classification deterministic approximate inference algorithms are widely used [16, 4, 5, 11].”
Online Kernel Principal Component Analysis: A Reduced-Order Model (2012, methodology): “Reducing the model order is a very active research area within the community of researchers in kernel machines [35], as well as Gaussian processes [36] [37, Chapter 4], since the underlying models of these methods have the form (2), as proven by the well-known Representer Theorem [25].”
Local Gaussian Process Regression for Real Time Online Model Learning (2008, methodology): “In SGP, the training data is approximated by a smaller set of so-called inducing inputs [2,5].”
Model Learning with Local Gaussian Process Regression (2009, methodology): “We compare the learning performance of LGP with the state-of-the-art in nonparametric regression, i.e., LWPR, ν-SVR, standard GPR and OGP in the context of approximating inverse robot dynamics.”
Estimation of the forgetting factor in kernel recursive least squares (2012, background): “On the other hand, KRLS algorithms are designed to perform regression in an online manner, i.e., when data points are made available on a one-at-a-time basis.”
Sparse Spectrum Gaussian Process Regression (2010, methodology): “Examples of such models are those proposed in Seeger et al. (2003), Smola and Bartlett (2001), Tresp (2000), Williams and Seeger (2001) and Csató and Opper (2002), as well as the Fully Independent Training Conditional (FITC) model, introduced as SparsePseudo-input GP (SPGP) by Snelson and…”
A Unifying Framework for Gaussian Process Pseudo-Point Approximations using Power Expectation Propagation (2016, methodology): “This begs the question: can the sparsification and the non-linear approximation be handled in a single EP inference stage, as for VFE? Astonishingly Csató and Opper not only developed such a method in 2002 (Csató and Opper, 2002), predating much of the work mentioned above, they showed that it is eq”