Multi-scale DeepOnet (Mscale-DeepOnet) for Mitigating Spectral Bias in Learning High Frequency Operators of Oscillatory Functions
By: Bo Wang, Lizuo Liu, Wei Cai
Potential Business Impact:
Helps computers understand wavy math problems better.
In this paper, a multi-scale DeepOnet (Mscale-DeepOnet) is proposed to reduce the spectral bias of the DeepOnet in learning high-frequency mapping between highly oscillatory functions, with an application to the nonlinear mapping between the coefficient of the Helmholtz equation and its solution. The Mscale-DeepOnet introduces the multiscale neural network in the branch and trunk networks of the original DeepOnet, the resulting Mscale-DeepOnet is shown to be able to capture various high-frequency components of the mapping itself and its image. Numerical results demonstrate the substantial improvement of the Mscale-DeepOnet for the problem of wave scattering in the high-frequency regime over the normal DeepOnet with a similar number of network parameters.
Similar Papers
Mitigating Spectral Bias in Neural Operators via High-Frequency Scaling for Physical Systems
Machine Learning (CS)
Fixes computer models that miss fast details.
Data-Efficient Deep Operator Network for Unsteady Flow: A Multi-Fidelity Approach with Physics-Guided Subsampling
Fluid Dynamics
Predicts fluid movement faster with less data.
Spectral Embedding via Chebyshev Bases for Robust DeepONet Approximation
Machine Learning (CS)
Solves hard math problems faster by understanding patterns.