Comparing recurrent and convolutional neural networks for predicting wave propagation

2020/02/20 by Stathi Fotiadis, Eduardo Pignatelli, Fotiadis, Stathi +9
Earth and Planetary Sciences · Engineering · #Advanced Fiber Optic Sensors #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Meteorological Phenomena and Simulations #Seismic Waves and Analysis

paper · pdf · doi:10.48550/arxiv.2002.08981

openalex publication_date 2020/02/20 · openalex created_date 2020/03/06 · openalex updated_date 2026/07/28

Abstract

Dynamical systems can be modelled by partial differential equations and numerical computations are used everywhere in science and engineering. In this work, we investigate the performance of recurrent and convolutional deep neural network architectures to predict the surface waves. The system is governed by the Saint-Venant equations. We improve on the long-term prediction over previous methods while keeping the inference time at a fraction of numerical simulations. We also show that convolutional networks perform at least as well as recurrent networks in this task. Finally, we assess the generalisation capability of each network by extrapolating in longer time-frames and in different physical settings.

Citations

• Unsupervised Learning of Video Representations using LSTMs
• Deep multi-scale video prediction beyond mean square error
• Physics Informed Deep Learning (Part I): Data-driven Solutions of Nonlinear Partial Differential Equations
• An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling
• PredRNN++: Towards A Resolution of the Deep-in-Time Dilemma in Spatiotemporal Predictive Learning
• Machine learning and the physical sciences
• FlowNet: Learning Optical Flow with Convolutional Networks
• Deep Learning Methods for Reynolds-Averaged Navier–Stokes Simulations of Airfoil Flows

Related

• Approximating the solution to wave propagation using deep neural networks
• Assessment of convolutional recurrent autoencoder network for learning wave propagation
• Rapid Phase-Resolved Prediction of Nonlinear Dispersive Waves Using\n Machine Learning
• On the reproducibility of fully convolutional neural networks for modeling time-space evolving physical systems
• Predicting Wave Dynamics using Deep Learning with Multistep Integration Inspired Attention and Physics-Based Loss Decomposition