no code implementations • 2 Sep 2020 • Enrui Zhang, Minglang Yin, George Em. Karniadakis
We apply Physics-Informed Neural Networks (PINNs) for solving identification problems of nonhomogeneous materials.
no code implementations • 24 Aug 2020 • Xiaoli Chen, Liu Yang, Jinqiao Duan, George Em. Karniadakis
The Fokker-Planck (FP) equation governing the evolution of the probability density function (PDF) is applicable to many disciplines but it requires specification of the coefficients for each case, which can be functions of space-time and not just constants, hence requiring the development of a data-driven modeling approach.
no code implementations • 5 Aug 2020 • Liu Yang, Constantinos Daskalakis, George Em. Karniadakis
Particle coordinates at a single time instant, possibly noisy or truncated, are recorded in each snapshot but are unpaired across the snapshots.
no code implementations • 7 May 2020 • Khemraj Shukla, Patricio Clark Di Leoni, James Blackshire, Daniel Sparkman, George Em. Karniadakis
The ultrasonic surface wave data is represented as a surface deformation on the top surface of a metal plate, measured by using the method of laser vibrometry.
no code implementations • 3 Apr 2020 • Yeonjong Shin, Jerome Darbon, George Em. Karniadakis
By adapting the Schauder approach and the maximum principle, we show that the sequence of minimizers strongly converges to the PDE solution in $C^0$.
no code implementations • 13 Mar 2020 • Liu Yang, Xuhui Meng, George Em. Karniadakis
In this Bayesian framework, the Bayesian neural network (BNN) combined with a PINN for PDEs serves as the prior while the Hamiltonian Monte Carlo (HMC) or the variational inference (VI) could serve as an estimator of the posterior.
1 code implementation • 11 Mar 2020 • Ehsan Kharazmi, Zhongqiang Zhang, George Em. Karniadakis
We formulate a general framework for hp-variational physics-informed neural networks (hp-VPINNs) based on the nonlinear approximation of shallow and deep neural networks and hp-refinement via domain decomposition and projection onto space of high-order polynomials.
1 code implementation • 6 Mar 2020 • Dixia Fan, Liu Yang, Michael S. Triantafyllou, George Em. Karniadakis
We demonstrate experimentally the feasibility of applying reinforcement learning (RL) in flow control problems by automatically discovering active control strategies without any prior knowledge of the flow physics.
Fluid Dynamics Robotics
1 code implementation • 11 Jan 2020 • Pengzhan Jin, Zhen Zhang, Aiqing Zhu, Yifa Tang, George Em. Karniadakis
We propose new symplectic networks (SympNets) for identifying Hamiltonian systems from data based on a composition of linear, activation and gradient modules.
1 code implementation • 2 Dec 2019 • Yuyao Chen, Lu Lu, George Em. Karniadakis, Luca Dal Negro
In this paper we employ the emerging paradigm of physics-informed neural networks (PINNs) for the solution of representative inverse scattering problems in photonic metamaterials and nano-optics technologies.
Computational Physics Optics
4 code implementations • 8 Oct 2019 • Lu Lu, Pengzhan Jin, George Em. Karniadakis
This universal approximation theorem is suggestive of the potential application of neural networks in learning nonlinear operators from data.
no code implementations • 25 Sep 2019 • Ameya D. Jagtap, Kenji Kawaguchi, George Em. Karniadakis
Furthermore, the proposed methods with the slope recovery are shown to accelerate the training process.
no code implementations • 23 Sep 2019 • Xuhui Meng, Zhen Li, Dongkun Zhang, George Em. Karniadakis
Consequently, compared to the original PINN approach, the proposed PPINN approach may achieve a significant speedup for long-time integration of PDEs, assuming that the CG solver is fast and can provide reasonable predictions of the solution, hence aiding the PPINN solution to converge in just a few iterations.
no code implementations • 19 Sep 2019 • Qiang Zheng, Lingzao Zeng, Zhendan Cao, George Em. Karniadakis
A fundamental problem in geostatistical modeling is to infer the heterogeneous geological field based on limited measurements and some prior spatial statistics.
no code implementations • 29 Aug 2019 • Liu Yang, George Em. Karniadakis
We propose a potential flow generator with $L_2$ optimal transport regularity, which can be easily integrated into a wide range of generative models including different versions of GANs and flow-based models.
no code implementations • 23 Jul 2019 • Yeonjong Shin, George Em. Karniadakis
In order to quantify the trainability, we study the probability distribution of the number of active neurons at the initialization.
1 code implementation • 27 May 2019 • Pengzhan Jin, Lu Lu, Yifa Tang, George Em. Karniadakis
To derive a meaningful bound, we study the generalization error of neural networks for classification problems in terms of data distribution and neural network smoothness.
no code implementations • 3 May 2019 • Dongkun Zhang, Ling Guo, George Em. Karniadakis
One of the open problems in scientific computing is the long-time integration of nonlinear stochastic partial differential equations (SPDEs).
no code implementations • 15 Mar 2019 • Lu Lu, Yeonjong Shin, Yanhui Su, George Em. Karniadakis
Numerical examples are provided to demonstrate the effectiveness of the new initialization procedure.
2 code implementations • 26 Feb 2019 • Xuhui Meng, George Em. Karniadakis
It is comprised of three NNs, with the first NN trained using the low-fidelity data and coupled to two high-fidelity NNs, one with activation functions and another one without, in order to discover and exploit nonlinear and linear correlations, respectively, between the low-fidelity and the high-fidelity data.
Computational Physics
no code implementations • 5 Nov 2018 • Liu Yang, Dongkun Zhang, George Em. Karniadakis
We developed a new class of physics-informed generative adversarial networks (PI-GANs) to solve in a unified manner forward, inverse and mixed stochastic problems based on a limited number of scattered measurements.
no code implementations • 27 Oct 2018 • Zhiping Mao, Zhen Li, George Em. Karniadakis
Instead of specifying the fPDEs with an ad hoc fractional order for nonlocal flocking dynamics, we learn the effective nonlocal influence function in fPDEs directly from particle trajectories generated by the agent-based simulations.
no code implementations • 21 Sep 2018 • Dongkun Zhang, Lu Lu, Ling Guo, George Em. Karniadakis
Here, we propose a new method with the objective of endowing the DNN with uncertainty quantification for both sources of uncertainty, i. e., the parametric uncertainty and the approximation uncertainty.
1 code implementation • 26 Aug 2018 • Maziar Raissi, Zhicheng Wang, Michael S. Triantafyllou, George Em. Karniadakis
Of interest is the prediction of the lift and drag forces on the structure given some limited and scattered information on the velocity field.
1 code implementation • ICLR 2019 • Lu Lu, Yanhui Su, George Em. Karniadakis
However, here we show that even for such activation, deep and narrow neural networks (NNs) will converge to erroneous mean or median states of the target function depending on the loss with high probability.
1 code implementation • 13 Aug 2018 • Maziar Raissi, Alireza Yazdani, George Em. Karniadakis
We present hidden fluid mechanics (HFM), a physics informed deep learning framework capable of encoding an important class of physical laws governing fluid motions, namely the Navier-Stokes equations.
2 code implementations • 4 Jan 2018 • Maziar Raissi, Paris Perdikaris, George Em. Karniadakis
The process of transforming observed data into predictive mathematical models of the physical world has always been paramount in science and engineering.
29 code implementations • 28 Nov 2017 • Maziar Raissi, Paris Perdikaris, George Em. Karniadakis
We introduce physics informed neural networks -- neural networks that are trained to solve supervised learning tasks while respecting any given law of physics described by general nonlinear partial differential equations.
23 code implementations • 28 Nov 2017 • Maziar Raissi, Paris Perdikaris, George Em. Karniadakis
We introduce physics informed neural networks -- neural networks that are trained to solve supervised learning tasks while respecting any given law of physics described by general nonlinear partial differential equations.
1 code implementation • 2 Aug 2017 • Maziar Raissi, George Em. Karniadakis
While there is currently a lot of enthusiasm about "big data", useful data is usually "small" and expensive to acquire.
1 code implementation • 29 Mar 2017 • Maziar Raissi, Paris Perdikaris, George Em. Karniadakis
Numerical Gaussian processes, by construction, are designed to deal with cases where: (1) all we observe are noisy data on black-box initial conditions, and (2) we are interested in quantifying the uncertainty associated with such noisy data in our solutions to time-dependent partial differential equations.
2 code implementations • 10 Jan 2017 • Maziar Raissi, George Em. Karniadakis
This work leverages recent advances in probabilistic machine learning to discover conservation laws expressed by parametric linear equations.
1 code implementation • 16 Jul 2016 • Maziar Raissi, Paris Perdikaris, George Em. Karniadakis
For more than two centuries, solutions of differential equations have been obtained either analytically or numerically based on typically well-behaved forcing and boundary conditions for well-posed problems.