Gridded Transformer Neural Processes for Large Unstructured Spatio-Temporal Data

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Ashman, Matthew, Diaconu, Cristiana, Langezaal, Eric, Weller, Adrian, Turner, Richard E.
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866910643557236736
author Ashman, Matthew
Diaconu, Cristiana
Langezaal, Eric
Weller, Adrian
Turner, Richard E.
author_facet Ashman, Matthew
Diaconu, Cristiana
Langezaal, Eric
Weller, Adrian
Turner, Richard E.
contents Many important problems require modelling large-scale spatio-temporal datasets, with one prevalent example being weather forecasting. Recently, transformer-based approaches have shown great promise in a range of weather forecasting problems. However, these have mostly focused on gridded data sources, neglecting the wealth of unstructured, off-the-grid data from observational measurements such as those at weather stations. A promising family of models suitable for such tasks are neural processes (NPs), notably the family of transformer neural processes (TNPs). Although TNPs have shown promise on small spatio-temporal datasets, they are unable to scale to the quantities of data used by state-of-the-art weather and climate models. This limitation stems from their lack of efficient attention mechanisms. We address this shortcoming through the introduction of gridded pseudo-token TNPs which employ specialised encoders and decoders to handle unstructured observations and utilise a processor containing gridded pseudo-tokens that leverage efficient attention mechanisms. Our method consistently outperforms a range of strong baselines on various synthetic and real-world regression tasks involving large-scale data, while maintaining competitive computational efficiency. The real-life experiments are performed on weather data, demonstrating the potential of our approach to bring performance and computational benefits when applied at scale in a weather modelling pipeline.
format Preprint
id arxiv_https___arxiv_org_abs_2410_06731
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Gridded Transformer Neural Processes for Large Unstructured Spatio-Temporal Data
Ashman, Matthew
Diaconu, Cristiana
Langezaal, Eric
Weller, Adrian
Turner, Richard E.
Machine Learning
Many important problems require modelling large-scale spatio-temporal datasets, with one prevalent example being weather forecasting. Recently, transformer-based approaches have shown great promise in a range of weather forecasting problems. However, these have mostly focused on gridded data sources, neglecting the wealth of unstructured, off-the-grid data from observational measurements such as those at weather stations. A promising family of models suitable for such tasks are neural processes (NPs), notably the family of transformer neural processes (TNPs). Although TNPs have shown promise on small spatio-temporal datasets, they are unable to scale to the quantities of data used by state-of-the-art weather and climate models. This limitation stems from their lack of efficient attention mechanisms. We address this shortcoming through the introduction of gridded pseudo-token TNPs which employ specialised encoders and decoders to handle unstructured observations and utilise a processor containing gridded pseudo-tokens that leverage efficient attention mechanisms. Our method consistently outperforms a range of strong baselines on various synthetic and real-world regression tasks involving large-scale data, while maintaining competitive computational efficiency. The real-life experiments are performed on weather data, demonstrating the potential of our approach to bring performance and computational benefits when applied at scale in a weather modelling pipeline.
title Gridded Transformer Neural Processes for Large Unstructured Spatio-Temporal Data
topic Machine Learning
url https://arxiv.org/abs/2410.06731