Fetching the paper…

Reading the bibliography…

2022

Physics-Constrained Generative Adversarial Networks for 3D Turbulence

Tretiak, Dima, Mohan, Arvind T., Livescu, Daniel

Original
PDF
Post on X

Understand

Generative Adversarial Networks (GANs) have received wide acclaim among the machine learning (ML) community for their ability to generate realistic 2D images.

ML is being applied more often to complex problems beyond those of computer vision.
However, current frameworks often serve as black boxes and lack physics embeddings, leading to poor ability in enforcing constraints and unreliable models.
In this work, we develop physics embeddings that can be stringently imposed, referred to as hard constraints, in the neural network architecture.

Physics-Constrained Generative Adversarial Networks for 3D Turbulence · Around

Built on

“Stabilizing Generative Adversarial Networks: A Survey”
Original
Maciej Wiatrak, Stefano. Albrecht and Andrew Nystrom · 1910
Earlier work this paper cites.
“Towards Physics-informed Deep Learning for Turbulent Flow Prediction”
Original
Rui Wang et al · 1911
Earlier work this paper cites.
“Analyzing and Improving the Image Quality of StyleGAN”
Original
Tero Karras et al · 1912
Earlier work this paper cites.
“Compact Finite Difference Schemes with Spectral-like Resolution”
S.. Lele · 1992
Earlier work this paper cites.
“Convolutional networks for images, speech, and time series”
Yann LeCun and Yoshua Bengio · 1995
Earlier work this paper cites.
“Lagrangian tetrad dynamics and the phenomenology of turbulence”
Michael Chertkov, Alain Pumir and Boris Shraiman · 1999
Earlier work this paper cites.
“Passive scalar wake behind a line source in grid turbulence”
Daniel Livescu, F.A. Jaberi and C.K. Madnia · 2000
Earlier work this paper cites.
“Turbulent flows”
Stephen Pope and Stephen Pope · 2000
Earlier work this paper cites.
“Embedding Hard Physical Constraints in Neural Network Coarse-Graining of 3D Turbulence”
Original
Arvind. Mohan, Nicholas Lubbers, Daniel Livescu and Michael Chertkov · 2002
Earlier work this paper cites.
“Linearly forced isotropic turbulence”
T.S. Lundgren · 2003
Earlier work this paper cites.
“Fourier Neural Operator for Parametric Partial Differential Equations”
Original
Zongyi Li et al · 2010
Earlier work this paper cites.
“Direct numerical simulations of Rayleigh-Taylor instability”
Daniel Livescu, Tie Wei and MR Petersen · 2011
Earlier work this paper cites.

Similar

“Generative Adversarial Networks”
Original
Ian. Goodfellow et al · 2014
Cited alongside, same era.
“Adam: A Method for Stochastic Optimization”
Original
Diederik. Kingma and Jimmy Ba · 2014
Cited alongside, same era.
“Unsupervised representation learning with deep convolutional generative adversarial networks”
Original
Alec Radford, Luke Metz and Soumith Chintala · 2015
Cited alongside, same era.
“Cyclical Learning Rates for Training Neural Networks”
Original
Leslie. Smith · 2015
Cited alongside, same era.
“Machine Learning Methods for Data-Driven Turbulence Modeling”
Ze Zhang and Karthikeyan Duraisamy · 2015
Cited alongside, same era.

Then

“Reaction analogy based forcing for incompressible scalar turbulence”
Don Daniel, Daniel Livescu and Jaiyoung Ryu · 2018
Later among the works it cites.
“From Deep to Physics-Informed Learning of Turbulence: Diagnostics”
Original
Ryan King, Oliver Hennigh, Arvind Mohan and Michael Chertkov · 2018
Later among the works it cites.
“Physics-informed neural networks: A deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations”
M. Raissi, P. Perdikaris and G.E. Karniadakis · 2018
Later among the works it cites.
“tempoGAN: A Temporally Coherent, Volumetric GAN for Super-resolution Fluid Flow”
Original
You Xie, Erik Franz, Mengyu Chu and Nils Thuerey · 2018
Later among the works it cites.
“Variational physics-informed neural networks for solving partial differential equations”
Original

Beyond the bibliography

alphaXiv searches the wider corpus for related work and actual follow-ups.

Open on alphaXiv

alphaXiv is searching for related work…

“Reynolds averaged turbulence modelling using deep neural networks with embedded invariance”

Julia Ling, Andrew Kurzawski and Jeremy Templeton · 2016

Cited alongside, same era.

“Deconvolution and Checkerboard Artifacts”

Augustus Odena, Vincent Dumoulin and Chris Olah · 2016

Cited alongside, same era.

“Improved Techniques for Training GANs”

Tim Salimans et al · 2016

Cited alongside, same era.

“Automatic Differentiation in PyTorch”

Adam Paszke et al · 2017

Cited alongside, same era.

“Attention Is All You Need”

Ashish Vaswani et al · 2017

Cited alongside, same era.

“A deep learning framework for turbulence modeling using data assimilation and feature extraction”

Atieh Alizadeh Moghaddam and Amir Sadaghiyani · 2018

Cited alongside, same era.

“Large Scale GAN Training for High Fidelity Natural Image Synthesis”

Original

Andrew Brock, Jeff Donahue and Karen Simonyan · 2018

Cited alongside, same era.

Ehsan Kharazmi, Zhongqiang Zhang and George Karniadakis · 2019

Later among the works it cites.

“Compressed convolutional LSTM: An efficient deep learning framework to model high fidelity 3D turbulence”

Arvind Mohan, Don Daniel, Michael Chertkov and Daniel Livescu · 2019

Later among the works it cites.

“fPINNs: Fractional physics-informed neural networks”

Guofei Pang, Lu Lu and George Karniadakis · 2019

Later among the works it cites.

“Enforcing statistical constraints in generative adversarial networks for modeling chaotic dynamical systems”

Jin-Long Wu et al · 2019

Later among the works it cites.

“Array programming with NumPy”

Charles. Harris et al · 2020

Later among the works it cites.

“Enforcing Physical Constraints in {CNN}s through Differentiable {PDE} Layer”

Chiyu”Max” Jiang, Karthik Kashinath, Prabhat and Philip Marcus · 2020

Later among the works it cites.

“Spatio-temporal deep learning models of 3D turbulence with physics informed diagnostics”

Arvind. Mohan, Dima Tretiak, Misha Chertkov and Daniel Livescu · 2020

Later among the works it cites.

“Physics-informed neural networks (PINNs) for fluid mechanics: A review”

Shengze Cai et al · 2022

Closest in time.