-
OceanNet: A principled neural operator-based digital twin for regional oceans
Authors:
Ashesh Chattopadhyay,
Michael Gray,
Tianning Wu,
Anna B. Lowe,
Ruoying He
Abstract:
While data-driven approaches demonstrate great potential in atmospheric modeling and weather forecasting, ocean modeling poses distinct challenges due to complex bathymetry, land, vertical structure, and flow non-linearity. This study introduces OceanNet, a principled neural operator-based digital twin for ocean circulation. OceanNet uses a Fourier neural operator and predictor-evaluate-corrector…
▽ More
While data-driven approaches demonstrate great potential in atmospheric modeling and weather forecasting, ocean modeling poses distinct challenges due to complex bathymetry, land, vertical structure, and flow non-linearity. This study introduces OceanNet, a principled neural operator-based digital twin for ocean circulation. OceanNet uses a Fourier neural operator and predictor-evaluate-corrector integration scheme to mitigate autoregressive error growth and enhance stability over extended time scales. A spectral regularizer counteracts spectral bias at smaller scales. OceanNet is applied to the northwest Atlantic Ocean western boundary current (the Gulf Stream), focusing on the task of seasonal prediction for Loop Current eddies and the Gulf Stream meander. Trained using historical sea surface height (SSH) data, OceanNet demonstrates competitive forecast skill by outperforming SSH predictions by an uncoupled, state-of-the-art dynamical ocean model forecast, reducing computation by 500,000 times. These accomplishments demonstrate the potential of physics-inspired deep neural operators as cost-effective alternatives to high-resolution numerical ocean models.
△ Less
Submitted 1 October, 2023;
originally announced October 2023.
-
GHM Wavelet Transform for Deep Image Super Resolution
Authors:
Ben Lowe,
Hadi Salman,
Justin Zhan
Abstract:
The GHM multi-level discrete wavelet transform is proposed as preprocessing for image super resolution with convolutional neural networks. Previous works perform analysis with the Haar wavelet only. In this work, 37 single-level wavelets are experimentally analyzed from Haar, Daubechies, Biorthogonal, Reverse Biorthogonal, Coiflets, and Symlets wavelet families. All single-level wavelets report si…
▽ More
The GHM multi-level discrete wavelet transform is proposed as preprocessing for image super resolution with convolutional neural networks. Previous works perform analysis with the Haar wavelet only. In this work, 37 single-level wavelets are experimentally analyzed from Haar, Daubechies, Biorthogonal, Reverse Biorthogonal, Coiflets, and Symlets wavelet families. All single-level wavelets report similar results indicating that the convolutional neural network is invariant to choice of wavelet in a single-level filter approach. However, the GHM multi-level wavelet achieves higher quality reconstructions than the single-level wavelets. Three large data sets are used for the experiments: DIV2K, a dataset of textures, and a dataset of satellite images. The approximate high resolution images are compared using seven objective error measurements. A convolutional neural network based approach using wavelet transformed images has good results in the literature.
△ Less
Submitted 16 April, 2022;
originally announced April 2022.
-
Proceedings 8th International Workshop on Developments in Computational Models
Authors:
Benedikt Löwe,
Glynn Winskel
Abstract:
The aim of the workshop series Developments in Computational Models (DCM) is to bring together researchers who are currently developing new computational models or new features for traditional computational models, in order to foster their interaction, to provide a forum for presenting new ideas and work in progress, and to enable newcomers to learn about current activities in this area. The eight…
▽ More
The aim of the workshop series Developments in Computational Models (DCM) is to bring together researchers who are currently developing new computational models or new features for traditional computational models, in order to foster their interaction, to provide a forum for presenting new ideas and work in progress, and to enable newcomers to learn about current activities in this area. The eighth workshop in the series, DCM 2012, was part of the celebrations of the Turing Centenary and was held as a satellite event of the Turing centenary conference Computability in Europe 2012 (CiE 2012) in Cambridge. It took place at Corpus Christi College in Cambridge on Sunday, 17 June 2013.
This electronic proceedings volume includes one of the keynote papers as well as revised versions of papers accepted for presentation by the programme committee.
△ Less
Submitted 28 March, 2014;
originally announced March 2014.
