MULTIGRID-AUGMENTED DEEP LEARNING PRECONDITIONERS FOR THE HELMHOLTZ EQUATION

Keyphrases

• Deep Learning 100%
• Convolutional Neural Network 100%
• Multigrid 100%
• Helmholtz Equation 100%
• U-Net 100%
• Two-dimensional Problems 33%
• Encoder 33%
• Data-driven Approach 33%
• Inverse Problem 33%
• Coarse Grid 33%
• Mini 33%
• General Set 33%
• Iterative Solvers 33%
• Training Data 33%
• Deflation 33%
• Data Augmentation 33%
• Generalization Ability 33%
• Shifted Laplacian 33%
• Training Scheme 33%
• Encoder Networks 33%
• Context Vector 33%
• Krylov Solvers 33%
• V-cycle 33%
• Helmholtz Equation at High Wavenumber 33%

Computer Science

• Deep Learning 100%
• Preconditioner 100%
• Convolutional Neural Network 60%
• U-Net 60%
• Driven Approach 20%
• Inverse Problem 20%
• Laplace Operator 20%
• Dimensional Problem 20%
• Data Augmentation 20%
• Generalization Ability 20%

Mathematics

• Deep Learning 100%
• Convolutional Neural Network 100%
• Helmholtz Equation 100%
• Residuals 33%
• Dimensional Problem 33%
• Laplace Operator 33%