Multi-fidelity experimental design for ice-sheet simulation

Pierre Thodoroff; Markus Kaiser; Rosie Williams; Robert Arthern; Scott Hosking; Neil D. Lawrence; James Byrne; Ieva Kazlauskaite

Back to publications

Multi-fidelity experimental design for ice-sheet simulation

Pierre Thodoroff, Markus Kaiser, Rosie Williams, Robert Arthern, Scott Hosking, Neil D. Lawrence, James Byrne, Ieva Kazlauskaite

, 2023.

Abstract

Computer simulations are becoming an essential tool in many scientific fields from molecular dynamics to aeronautics. In glaciology, future predictions of sea level change require input from ice sheet models. Due to uncertainties in the forcings and the parameter choices for such models, many different realisations of the model are needed in order to produce probabilistic forecasts of sea level change. For these reasons, producing robust probabilistic forecasts from an ensemble of model simulations over regions of interest can be extremely expensive for many ice sheet models. Multi-fidelity experimental design (MFED) is a strategy that models the high-fidelity output of the simulator by combining information from various resolutions in an attempt to minimize the computational costs of the process and maximize the accuracy of the posterior. In this paper, we present an application of MFED to an ice-sheet simulator and demonstrate potential computational savings by modelling the relationship between spatial resolutions. We also analyze the behavior of MFED strategies using theoretical results from sub-modular maximization.

Links

Cite this Paper

BibTeX


@Article{multifidelity-experimental-design-for-ice-sheet-simulation,
  title = 	 {Multi-fidelity experimental design for ice-sheet simulation},
  author = 	 {Thodoroff, Pierre and Kaiser, Markus and Williams, Rosie and Arthern, Robert and Hosking, Scott and Lawrence, Neil D. and Byrne, James and Kazlauskaite, Ieva},
  year = 	 {2023},
  url = 	 {/publications/multifidelity-experimental-design-for-ice-sheet-simulation.html},
  abstract = 	 {Computer simulations are becoming an essential tool in many scientific fields 
from molecular dynamics to aeronautics. In glaciology, future predictions of 
sea level change require input from ice sheet models. Due to uncertainties in   the forcings and the parameter choices for such models, many different 
realisations of the model are needed in order to produce probabilistic 
forecasts of sea level change. For these reasons, producing robust 
probabilistic forecasts from an ensemble of model simulations over regions 
of interest can be extremely expensive for many ice sheet models. 
Multi-fidelity experimental design (MFED) is a strategy that models the 
high-fidelity output of the simulator by combining information from various 
resolutions in an attempt to minimize the computational costs of the process 
and maximize the accuracy of the posterior. In this paper, we present an 
application of MFED to an ice-sheet simulator and demonstrate potential 
computational savings by modelling the relationship between spatial 
resolutions. We also analyze the behavior of MFED strategies using 
theoretical results from sub-modular maximization.
}
}

Endnote

%0 Journal Article
%T Multi-fidelity experimental design for ice-sheet simulation
%A Pierre Thodoroff
%A Markus Kaiser
%A Rosie Williams
%A Robert Arthern
%A Scott Hosking
%A Neil D. Lawrence
%A James Byrne
%A Ieva Kazlauskaite
%D 2023	
%F multifidelity-experimental-design-for-ice-sheet-simulation
%U /publications/multifidelity-experimental-design-for-ice-sheet-simulation.html
%X Computer simulations are becoming an essential tool in many scientific fields 
from molecular dynamics to aeronautics. In glaciology, future predictions of 
sea level change require input from ice sheet models. Due to uncertainties in   the forcings and the parameter choices for such models, many different 
realisations of the model are needed in order to produce probabilistic 
forecasts of sea level change. For these reasons, producing robust 
probabilistic forecasts from an ensemble of model simulations over regions 
of interest can be extremely expensive for many ice sheet models. 
Multi-fidelity experimental design (MFED) is a strategy that models the 
high-fidelity output of the simulator by combining information from various 
resolutions in an attempt to minimize the computational costs of the process 
and maximize the accuracy of the posterior. In this paper, we present an 
application of MFED to an ice-sheet simulator and demonstrate potential 
computational savings by modelling the relationship between spatial 
resolutions. We also analyze the behavior of MFED strategies using 
theoretical results from sub-modular maximization.

RIS


TY  - JOUR
TI  - Multi-fidelity experimental design for ice-sheet simulation
AU  - Pierre Thodoroff
AU  - Markus Kaiser
AU  - Rosie Williams
AU  - Robert Arthern
AU  - Scott Hosking
AU  - Neil D. Lawrence
AU  - James Byrne
AU  - Ieva Kazlauskaite
DA  - 2023/07/17	
ID  - multifidelity-experimental-design-for-ice-sheet-simulation
UR  - /publications/multifidelity-experimental-design-for-ice-sheet-simulation.html
AB  - Computer simulations are becoming an essential tool in many scientific fields 
from molecular dynamics to aeronautics. In glaciology, future predictions of 
sea level change require input from ice sheet models. Due to uncertainties in   the forcings and the parameter choices for such models, many different 
realisations of the model are needed in order to produce probabilistic 
forecasts of sea level change. For these reasons, producing robust 
probabilistic forecasts from an ensemble of model simulations over regions 
of interest can be extremely expensive for many ice sheet models. 
Multi-fidelity experimental design (MFED) is a strategy that models the 
high-fidelity output of the simulator by combining information from various 
resolutions in an attempt to minimize the computational costs of the process 
and maximize the accuracy of the posterior. In this paper, we present an 
application of MFED to an ice-sheet simulator and demonstrate potential 
computational savings by modelling the relationship between spatial 
resolutions. We also analyze the behavior of MFED strategies using 
theoretical results from sub-modular maximization.

ER  -

APA


Thodoroff, P., Kaiser, M., Williams, R., Arthern, R., Hosking, S., Lawrence, N.D., Byrne, J. & Kazlauskaite, I.. (2023). Multi-fidelity experimental design for ice-sheet simulation.  Available from /publications/multifidelity-experimental-design-for-ice-sheet-simulation.html.