Multi-fidelity methods are prominently used when cheaply-obtained, but
possibly biased and noisy, observations must be effectively combined with
limited or expensive true data in order to construct reliable models. This
arises in both fundamental machine learning procedures such as Bayesian
optimization, as well as more practical science and engineering applications.
In this paper we develop a novel multi-fidelity model which treats layers of a
deep Gaussian process as fidelity levels, and uses a variational inference
scheme to propagate uncertainty across them. This allows for capturing
nonlinear correlations between fidelities with lower risk of overfitting than
existing methods exploiting compositional structure, which are conversely
burdened by structural assumptions and constraints. We show that the proposed
approach makes substantial improvements in quantifying and propagating
uncertainty in multi-fidelity set-ups, which in turn improves their
effectiveness in decision making pipelines.
Description
[1903.07320] Deep Gaussian Processes for Multi-fidelity Modeling
%0 Journal Article
%1 cutajar2019gaussian
%A Cutajar, Kurt
%A Pullin, Mark
%A Damianou, Andreas
%A Lawrence, Neil
%A González, Javier
%D 2019
%K gaussian-proceses
%T Deep Gaussian Processes for Multi-fidelity Modeling
%U http://arxiv.org/abs/1903.07320
%X Multi-fidelity methods are prominently used when cheaply-obtained, but
possibly biased and noisy, observations must be effectively combined with
limited or expensive true data in order to construct reliable models. This
arises in both fundamental machine learning procedures such as Bayesian
optimization, as well as more practical science and engineering applications.
In this paper we develop a novel multi-fidelity model which treats layers of a
deep Gaussian process as fidelity levels, and uses a variational inference
scheme to propagate uncertainty across them. This allows for capturing
nonlinear correlations between fidelities with lower risk of overfitting than
existing methods exploiting compositional structure, which are conversely
burdened by structural assumptions and constraints. We show that the proposed
approach makes substantial improvements in quantifying and propagating
uncertainty in multi-fidelity set-ups, which in turn improves their
effectiveness in decision making pipelines.
@article{cutajar2019gaussian,
abstract = {Multi-fidelity methods are prominently used when cheaply-obtained, but
possibly biased and noisy, observations must be effectively combined with
limited or expensive true data in order to construct reliable models. This
arises in both fundamental machine learning procedures such as Bayesian
optimization, as well as more practical science and engineering applications.
In this paper we develop a novel multi-fidelity model which treats layers of a
deep Gaussian process as fidelity levels, and uses a variational inference
scheme to propagate uncertainty across them. This allows for capturing
nonlinear correlations between fidelities with lower risk of overfitting than
existing methods exploiting compositional structure, which are conversely
burdened by structural assumptions and constraints. We show that the proposed
approach makes substantial improvements in quantifying and propagating
uncertainty in multi-fidelity set-ups, which in turn improves their
effectiveness in decision making pipelines.},
added-at = {2019-03-22T13:19:30.000+0100},
author = {Cutajar, Kurt and Pullin, Mark and Damianou, Andreas and Lawrence, Neil and González, Javier},
biburl = {https://www.bibsonomy.org/bibtex/268175796729f4a2b23c9f1a8a96b04f8/kirk86},
description = {[1903.07320] Deep Gaussian Processes for Multi-fidelity Modeling},
interhash = {26f895f3f393a5c22ab3b542f6e7019e},
intrahash = {68175796729f4a2b23c9f1a8a96b04f8},
keywords = {gaussian-proceses},
note = {cite arxiv:1903.07320},
timestamp = {2019-03-22T13:19:30.000+0100},
title = {Deep Gaussian Processes for Multi-fidelity Modeling},
url = {http://arxiv.org/abs/1903.07320},
year = 2019
}