A Bayesian Committee Machine

By Volker Tresp. Neural Computation, 2000

475 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Control and Systems Engineering

View Paper: https://www.semanticscholar.org/paper/7eeea351d07ef9cc2d572d857e944ab59ee7c61b

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

Citations in the corpus

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 M 2 N where M N [1] [2] [3] [4] [5] [6] [7] [8] [9].”
Distributed Gaussian Processes (2015, methodology): “…amongst individual computing units and propose the generalised BCM (gBMC), a new family of hierarchical PoE-GP models that (i) includes the BCM (Tresp, 2000) and to some degree the generalised PoE-GP (Cao & Fleet, 2014) as special cases, (ii) provides consistent approximations of a full GP,…”
Algorithm runtime prediction: Methods & evaluation (2012, methodology): “We experimented with the Bayesian committee machine [110], the informative vector machine [72], and the projected process (PP) approximation [94].”
Large-scale kernel machines (2007, background): “There are many strategies for specific problems which try to reduce this computational complexity by searching for a sparse representation of the data (Williams and Seeger, 2001; Smola and Bartlett, 2001; Fine and Scheinberg, 2001; Lee and Mangasarian, 2001; Lawrence et al., 2003; Csato and Opper, 2”
A Parallel Mixture of SVMs for Very Large Scale Problems (2001, methodology): “Finally, the Bayesian Committee Machine [8] is a technique to partition the data into several subsets, train SVMs on the individual subsets and then use a specific combination scheme based on the covariance of the test data to combine the predictions.”
Fast and Accurate Uncertainty Estimation in Chemical Machine Learning. (2018, background): “59 It also suggests the possibility of developing committee models 36,60 in which multiple ML predictions are performed, diﬀering in the deﬁnition of the kernel and/or regularization.”
Artificial Neural Networks — ICANN 2001 (2001, methodology): “Architecture selection is a long studied question for MLPs, and many approaches, such as bootstrapping, pruning, regularization, stopped training or information criteria have been proposed (see [3,6,8] for an overview of these approaches).”
Sparse Spectrum Gaussian Process Regression (2010, methodology): “One of the main practical limitations of Gaussian processes (GPs) for machinelear ing (Rasmussen and Williams, 2006) is that in a direct implementation the computational and memory requirements scale asO(n2) andO(n3), respectively.”
Gaussian processes:iterative sparse approximations (2002, methodology): “2) given by the representer theorem of Kimeldorf and Wahba [1971] (generalised by Schölkopf et al.”
Recent advances in surrogate-based optimization (2009, background): “Ensembles or committees (as they are known in the machine learning literature [79]) are a powerful concept, indeed blind Kriging could be viewed as form of committee model.”
Approximation Methods for Gaussian Process Regression (2007, methodology): “A reasonable choice, also recommended by Tresp (2000) may be to choose k = n/m blocks, each of size m × m.”
Dynamic systems identification with Gaussian processes (2005, methodology): “These attempts include the use of the Nystro¨m method [16], selection mechanisms [17], mixtures of GPs [18] and the Bayesian committee machine [19].”
Handbook of Neural Network Signal Processing (2000, methodology): “For instance, the kernel-Adatron [39] is derived from the Adatron algorithm originally proposed by Anlauf and Biehl [3] in a statistical mechanics setting.”
Cautious Model Predictive Control Using Gaussian Process Regression (2017, background): “, heuristically, as a subset of the original data points, by treating them as hyperparameters and optimizing over their location [14], or letting them coincide with test points [15], which is often referred to as transductive learning.”
Policy committee for adaptation in multi-domain spoken dialogue systems (2015, methodology): “Another advantage of MBCM is that it does not require storing a separate generic policy model but only ever produces in-domain models that have the ability to contribute to action-selection for other domains.”
Fast Forward Selection to Speed Up Sparse Gaussian Process Regression (2003, methodology): “A range of sparse GP inference approximations have addressed this problem (see Smola and Bartlett, 2001, Williams and Seeger, 2001, Csató and Opper, 2002, Tresp, 2000, Tipping, 2001).”
Transformations of Gaussian Process Priors (2004, methodology): “, 2002), and Bayesian committee machine (Tresp, 2000).”
Large Scale Machine Learning (2004, methodology): “In Chapter 3 we introduced Support Vector Machines (SVMs), a recent and important machine learning algorithm for classification and regression proposed by Vapnik (1995). We showed that SVM training requires the minimization of a quadratic problem under constraints.”