Summary
2023
Session Number:B4L-3
Session:
Number:B4L-33
Super Resolution of Numerical Solutions of Nonlinear Elliptic Equations by DeepONet
Chen Yuhan, Matsubara Takashi, Yaguchi Takaharu,
pp.370-373
Publication Date:2023-09-21
Online ISSN:2188-5079
DOI:10.34385/proc.76.B4L-33
PDF download (284KB)
Summary:
In recent years, super resolution has been widely used to enhance the resolution of images, thereby obtaining clearer images. In the field of physics simulation, super-resolution technique has also been actively developed. However, existing methods based on standard neural networks work on a single scale factor, making it unsuitable for cases where arbitrary super-resolution scale factors are required. Meanwhile, DeepONet is a deep learning framework capable of learning operators between function spaces. In this paper, we propose a DeepONet-based approach to super resolution of numerical solutions of nonlinear partial differential equations. Some experimental results are also provided.