Continuous Convolutional Neural Network

Continuous Convolutional Neural Networks (CCNNs) represent a novel approach to convolutional neural networks that operate on continuous rather than discrete data, enabling processing of data with varying resolutions and dimensions without architectural modifications. Current research focuses on developing CCNN architectures for diverse applications, including super-resolution imaging, fluid dynamics simulation, and prediction tasks in complex systems like nuclear fusion plasmas, often demonstrating superior performance and efficiency compared to traditional discrete CNNs. This flexibility and improved performance make CCNNs a promising tool across various scientific domains and engineering applications requiring efficient and adaptable image and signal processing.

Papers

December 3, 2023

Continuous Convolutional Neural Networks for Disruption Prediction in Nuclear Fusion Plasmas
William F Arnold, Lucas Spangher, Christina Rea
Plasma Dynamic DIII D Tokamak Disruption Prediction Decarbonization Study Continuous Convolutional Neural Network

November 13, 2023

Affine Invariance in Continuous-Domain Convolutional Neural Networks
Ali Mohaddes, Johannes Lederer
Affine Transformation Invariance Property Isometry Invariant Affine Transform Continuous Convolutional Neural Network

June 1, 2023

Spatio-Angular Convolutions for Super-resolution in Diffusion MRI
Matthew Lyon, Paul Armitage, Mauricio A Álvarez
Super Resolution Diffusion MRI Axial Resolution Spatial Convolution Continuous Convolutional Neural Network

January 25, 2023

Modelling Long Range Dependencies in $N$D: From Task-Specific to a General Purpose CNN
David M. Knigge, David W. Romero, Albert Gu, Efstratios Gavves, Erik J. Bekkers, Jakub M. Tomczak, Mark Hoogendoorn, Jan-Jakob Sonke
CNN Model Task Specific Convolutional Kernel Long Range Dependency Performant Neural Network Continuous Convolutional Neural Network

October 12, 2022

Guaranteed Conservation of Momentum for Learning Particle-based Fluid Dynamics
Lukas Prantl, Benjamin Ummenhofer, Vladlen Koltun, Nils Thuerey
Generalization Performance Residual Momentum Physical Simulation Environmental Conservation Continuous Convolutional Neural Network Vortex Particle

June 7, 2022

Towards a General Purpose CNN for Long Range Dependencies in $N$D
David W. Romero, David M. Knigge, Albert Gu, Erik J. Bekkers, Efstratios Gavves, Jakub M. Tomczak, Mark Hoogendoorn
CNN Architecture Convolutional Kernel Long Range Dependency Single Convolutional Continuous Convolutional Neural Network