Sparse On-Line Gaussian Processes

By Lehel Csató, Manfred Opper. 2002

851 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Control and Systems Engineering, Optimization & Training, 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 Control and Systems Engineering

In Optimization & Training

In Probabilistic & Bayesian Methods

Citations in the corpus

Continual Learning via Sequential Function-Space Variational Inference (2023, methodology): “…sparse GP regression (Cheng and Boots, 2016), streaming sparse GP s (Bui et al., 2017) and online sparse multi-output GP regression (Yang et al., 2019) built on the work of Csat´o and Opper (2002) and Csat´o (2002), and are effective approaches to continual learning for regression tasks.”
Variational Auto-Regressive Gaussian Processes for Continual Learning (2020, methodology): “Historically, Csató and Opper [7] and Csató [6] presented seminal works in the context of online regression.”
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…”
Gaussian Processes for Big Data (2013, background): “An alternative is to build a low rank approximation to the covariance matrix based around ‘inducing vari-ables’ [see e.g. Csat´o and Opper, 2002, Seeger et al., 2003, Qui˜nonero Candela and Rasmussen, 2005, Tit-sias, 2009].”
Variational Bayesian Multinomial Probit Regression with Gaussian Process Priors (2006, methodology): “Only binary classication has been considered in (Lawrence, et al 2003; Csato and Opper, 2002; Quinonero-Candela and Winther, 2003) and it is clear from (Seeger and Jordan, 2004) that extension of ADF based approximations such as IVM to the multi-class problem is not at all straightforward when a m”
Model learning for robot control: a survey (2011, methodology): “In such cases, the learned models need to be adapted to the changes and, thus, online model learning is necessary ( Csato and Opper 2002; Nguyen-Tuong and Peters 2009).”
Fast Forward Selection to Speed Up Sparse Gaussian Process Regression (2003, methodology): “The idea of approximating the likelihood using a linear projection from d variables is not restricted to using a d-subset of u (as has been noted in Smola and Bartlett (2001), Csató and Opper (2002)).”
A Bayesian spatiotemporal model for very large data sets (2010, positive): “These ideas have also been used to approximate Gaussian process models (Csato and Opper, 2002) and applied to very large data sets (Cornford et al., 2005).”
Learning to soar: Resource-constrained exploration in reinforcement learning (2015, methodology): “These were also compared against the performance of an e-greedy policy, a sparseeGP approach (Csató and Opper, 2002) using the full information value (that is, the first information measure) and a timestep-weighted policy, iGP-SARSA(l), adapted from Chung et al.”
A Practical and Conceptual Framework for Learning in Control (2010, methodology): “(1999), Smola and Bartlett (2001), Csató and Opper (2002), and Seeger et al. (2003), which are not discussed further in this report. Using the Γ-notation, the log-marginal likelihood of the sparse GP methods by Titsias (2009) and Snelson and Ghahramani (2006) and Snelson (2007) is given by”
Construction of approximation spaces for reinforcement learning (2013, background): “However, finding a suitable subset is not trivial (Smola and Schölkopf, 2000; Csató and Opper, 2002).”
MCMC for Variationally Sparse Gaussian Processes (2015, methodology): “Early strategies were based on retaining a sub-set of the data [2].”
Probabilistic Differential Dynamic Programming (2014, background): “[11][18][19], etc) may be used to reduce the computational cost of GP dynamics propagation.”
Analysis of Some Methods for Reduced Rank Gaussian Process Regression (2003, methodology): “These include the sparse iterative schemes of Csató (2002), Csató and Opper (2002), Seeger (2003), and Lawrence et al. (2003), all based R. Murray-Smith, R. Shorten (Eds.)”
When Gaussian Process Meets Big Data: A Review of Scalable GPs (2018, methodology): “For the complicated online regression, the model [65] should (i) have real-time adaptation to the streaming data; and (ii) handle large-scale case since the new data is continuously arriving.”