Environmental sensor placement with convolutional Gaussian neural processes
Abstract Environmental sensors are crucial for monitoring weather conditions and the impacts of climate change. However, it is challenging to place sensors in a way that maximises the informativeness of their measurements, particularly in remote regions like Antarctica. Probabilistic machine learnin...
Published in: | Environmental Data Science |
---|---|
Main Authors: | , , , , , , , , , , |
Other Authors: | , , , |
Format: | Article in Journal/Newspaper |
Language: | English |
Published: |
Cambridge University Press (CUP)
2023
|
Subjects: | |
Online Access: | http://dx.doi.org/10.1017/eds.2023.22 https://www.cambridge.org/core/services/aop-cambridge-core/content/view/S2634460223000225 |
id |
crcambridgeupr:10.1017/eds.2023.22 |
---|---|
record_format |
openpolar |
spelling |
crcambridgeupr:10.1017/eds.2023.22 2024-09-30T14:25:11+00:00 Environmental sensor placement with convolutional Gaussian neural processes Andersson, Tom R. Bruinsma, Wessel P. Markou, Stratis Requeima, James Coca-Castro, Alejandro Vaughan, Anna Ellis, Anna-Louise Lazzara, Matthew A. Jones, Dani Hosking, Scott Turner, Richard E. Alan Turing Institute National Science Foundation UK Research and Innovation Engineering and Physical Sciences Research Council Engineering and Physical Sciences Research Council 2023 http://dx.doi.org/10.1017/eds.2023.22 https://www.cambridge.org/core/services/aop-cambridge-core/content/view/S2634460223000225 en eng Cambridge University Press (CUP) http://creativecommons.org/licenses/by/4.0 Environmental Data Science volume 2 ISSN 2634-4602 journal-article 2023 crcambridgeupr https://doi.org/10.1017/eds.2023.22 2024-09-04T04:04:45Z Abstract Environmental sensors are crucial for monitoring weather conditions and the impacts of climate change. However, it is challenging to place sensors in a way that maximises the informativeness of their measurements, particularly in remote regions like Antarctica. Probabilistic machine learning models can suggest informative sensor placements by finding sites that maximally reduce prediction uncertainty. Gaussian process (GP) models are widely used for this purpose, but they struggle with capturing complex non-stationary behaviour and scaling to large datasets. This paper proposes using a convolutional Gaussian neural process (ConvGNP) to address these issues. A ConvGNP uses neural networks to parameterise a joint Gaussian distribution at arbitrary target locations, enabling flexibility and scalability. Using simulated surface air temperature anomaly over Antarctica as training data, the ConvGNP learns spatial and seasonal non-stationarities, outperforming a non-stationary GP baseline. In a simulated sensor placement experiment, the ConvGNP better predicts the performance boost obtained from new observations than GP baselines, leading to more informative sensor placements. We contrast our approach with physics-based sensor placement methods and propose future steps towards an operational sensor placement recommendation system. Our work could help to realise environmental digital twins that actively direct measurement sampling to improve the digital representation of reality. Article in Journal/Newspaper Antarc* Antarctica Cambridge University Press Environmental Data Science 2 |
institution |
Open Polar |
collection |
Cambridge University Press |
op_collection_id |
crcambridgeupr |
language |
English |
description |
Abstract Environmental sensors are crucial for monitoring weather conditions and the impacts of climate change. However, it is challenging to place sensors in a way that maximises the informativeness of their measurements, particularly in remote regions like Antarctica. Probabilistic machine learning models can suggest informative sensor placements by finding sites that maximally reduce prediction uncertainty. Gaussian process (GP) models are widely used for this purpose, but they struggle with capturing complex non-stationary behaviour and scaling to large datasets. This paper proposes using a convolutional Gaussian neural process (ConvGNP) to address these issues. A ConvGNP uses neural networks to parameterise a joint Gaussian distribution at arbitrary target locations, enabling flexibility and scalability. Using simulated surface air temperature anomaly over Antarctica as training data, the ConvGNP learns spatial and seasonal non-stationarities, outperforming a non-stationary GP baseline. In a simulated sensor placement experiment, the ConvGNP better predicts the performance boost obtained from new observations than GP baselines, leading to more informative sensor placements. We contrast our approach with physics-based sensor placement methods and propose future steps towards an operational sensor placement recommendation system. Our work could help to realise environmental digital twins that actively direct measurement sampling to improve the digital representation of reality. |
author2 |
Alan Turing Institute National Science Foundation UK Research and Innovation Engineering and Physical Sciences Research Council Engineering and Physical Sciences Research Council |
format |
Article in Journal/Newspaper |
author |
Andersson, Tom R. Bruinsma, Wessel P. Markou, Stratis Requeima, James Coca-Castro, Alejandro Vaughan, Anna Ellis, Anna-Louise Lazzara, Matthew A. Jones, Dani Hosking, Scott Turner, Richard E. |
spellingShingle |
Andersson, Tom R. Bruinsma, Wessel P. Markou, Stratis Requeima, James Coca-Castro, Alejandro Vaughan, Anna Ellis, Anna-Louise Lazzara, Matthew A. Jones, Dani Hosking, Scott Turner, Richard E. Environmental sensor placement with convolutional Gaussian neural processes |
author_facet |
Andersson, Tom R. Bruinsma, Wessel P. Markou, Stratis Requeima, James Coca-Castro, Alejandro Vaughan, Anna Ellis, Anna-Louise Lazzara, Matthew A. Jones, Dani Hosking, Scott Turner, Richard E. |
author_sort |
Andersson, Tom R. |
title |
Environmental sensor placement with convolutional Gaussian neural processes |
title_short |
Environmental sensor placement with convolutional Gaussian neural processes |
title_full |
Environmental sensor placement with convolutional Gaussian neural processes |
title_fullStr |
Environmental sensor placement with convolutional Gaussian neural processes |
title_full_unstemmed |
Environmental sensor placement with convolutional Gaussian neural processes |
title_sort |
environmental sensor placement with convolutional gaussian neural processes |
publisher |
Cambridge University Press (CUP) |
publishDate |
2023 |
url |
http://dx.doi.org/10.1017/eds.2023.22 https://www.cambridge.org/core/services/aop-cambridge-core/content/view/S2634460223000225 |
genre |
Antarc* Antarctica |
genre_facet |
Antarc* Antarctica |
op_source |
Environmental Data Science volume 2 ISSN 2634-4602 |
op_rights |
http://creativecommons.org/licenses/by/4.0 |
op_doi |
https://doi.org/10.1017/eds.2023.22 |
container_title |
Environmental Data Science |
container_volume |
2 |
_version_ |
1811643788437749760 |