Neural networks for wave propagation
Neural-Network, Wave Propagation, PINN, Fourier Neural Operator
The study of wave propagation lies at the core of applications across various fields of physics, ranging from quantum
theory to reservoir monitoring in geophysics. Although extensively investigated and applied over the past decades, several
challenges remain, such as the computational cost associated with simulating wave propagation in heterogeneous media,
handling irregular domains, and solving problems with corrupted data. These issues continue to be actively researched across
multiple scientific disciplines. In light of this, recent advances in machine learning—and more specifically, in neural
networks—have shown promise as powerful tools to assist with such complex tasks. In this work, we explore different
applications of wave propagation using physics-informed neural networks and neural operators, incorporating prior knowledge
about the properties of wave solutions in order to enhance the learning capabilities of the networks.