Search papers, labs, and topics across Lattice.
This paper introduces the Multi-Resolution Finite-Volume-inspired network (MuRFiV), which integrates physics-informed deep learning to enhance the prediction of complex spatiotemporal dynamics governed by partial differential equations. By embedding the conservative properties of finite volume methods into the neural network architecture, MuRFiV achieves superior long-term prediction accuracy and stability compared to traditional data-driven approaches. The results demonstrate that this hybrid framework effectively mitigates issues of error accumulation and generalizability in spatiotemporal predictions, making it a significant advancement in the field.
MuRFiV achieves unprecedented long-term prediction accuracy in spatiotemporal dynamics by merging finite-volume principles with deep learning, outperforming conventional neural networks.
Predicting complex spatiotemporal dynamics in physical processes often demands computationally expensive numerical methods or data-driven neural networks that suffer from high training costs, error accumulation, and limited generalizability to unseen parameters. An effective approach to address these challenges is leveraging physics priors in training neural networks, known as physics-informed deep learning (PiDL). In this work, we introduce the Multi-Resolution Finite-Volume-inspired network, MuRFiV, designed to capitalize on the conservative property of finite volume on the global scale and the expressive power of deep learning on the local scale. We demonstrate the effectiveness of MuRFiV on several spatio-temporal systems governed by partial differential equations (PDEs), including Burgers' equation, shallow water equations, and incompressible Navier-Stokes equations. By embedding PDE information into the deep learning architecture, MuRFiV achieves strong long-term prediction accuracy and remains stable over very long autoregressive rollouts, significantly outperforming data-driven neural network baselines. This result highlights the promise of combining multiresolution learning with finite-volume-inspired inductive bias for accurate and robust long-term prediction of complex dynamics.
