Deep learning models for acoustic scattering problems

Abstract
We develop deep learning (DL) models based on discriminative and generative networks to solve the forward and inverse acoustic scattering problems and show how these models streamlines the inverse design process by eliminating the degenerate solution space. Specifically, we present DL frameworks for designing broadband acoustic cloaks and arbitrarily-shape acoustic object recognition for underwater applications.

Acknowledgements
The work described in here is supported by King Abdullah University of Science and Technology (KAUST) Artificial Intelligence Initiative Fund and KAUST Baseline Research Fund No. BAS/1/1626-01-01.

Publisher
International Commission for Acoustics (ICA)

Conference/Event Name
24th International Congress on Acoustics, ICA 2022

Additional Links
https://ica2022korea.org/data/ICA2022_Program.pdf

Permanent link to this record