no code implementations • 28 Feb 2024 • Koichiro Yawata, Kai Fukami, Kunihiko Taira, Hiroya Nakao
We present a phase autoencoder that encodes the asymptotic phase of a limit-cycle oscillator, a fundamental quantity characterizing its synchronization dynamics.
1 code implementation • 13 May 2023 • Kai Fukami, Kunihiko Taira
We reveal that the nonlinear vortical flow field over time and parameter space can be compressed to only three variables with a lift-augmented autoencoder while holding the essence of the original high-dimensional physics.
no code implementations • 26 Jan 2023 • Kai Fukami, Koji Fukagata, Kunihiko Taira
This paper surveys machine-learning-based super-resolution reconstruction for vortical flows.
no code implementations • 16 Sep 2021 • Masaki Morimoto, Kai Fukami, Romit Maulik, Ricardo Vinuesa, Koji Fukagata
The average of such an ensemble can be regarded as the `mean estimation', whereas its standard deviation can be used to construct `confidence intervals', which enable us to perform uncertainty quantification (UQ) with regard to the training process of neural networks.
no code implementations • 7 Jan 2021 • Masaki Morimoto, Kai Fukami, Kai Zhang, Aditya G. Nair, Koji Fukagata
We then discuss the influence of various parameters and operations on the CNN performance, with the utilization of autoencoder (AE).
Dimensionality Reduction Fluid Dynamics Computational Physics
1 code implementation • 3 Jan 2021 • Kai Fukami, Romit Maulik, Nesar Ramachandra, Koji Fukagata, Kunihiko Taira
This reconstruction problem is especially difficult when sensors are sparsely positioned in a seemingly random or unorganized manner, which is often encountered in a range of scientific and engineering problems.
1 code implementation • 8 May 2020 • Romit Maulik, Kai Fukami, Nesar Ramachandra, Koji Fukagata, Kunihiko Taira
We consider the use of probabilistic neural networks for fluid flow surrogate modeling and data recovery.
Fluid Dynamics
1 code implementation • 17 Mar 2020 • Kazuto Hasegawa, Kai Fukami, Takaaki Murata, Koji Fukagata
The present ML-ROMs are trained on a set of 80 bluff body shapes and tested on a different set of 20 bluff body shapes not used for training, with both training and test shapes chosen from the same random distribution.
Fluid Dynamics Computational Physics