Laurent Demanet
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Laurent Demanet has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Hongyu Sun
Department of Earth, Environmental and Resource Sciences, The University of Texas at El Paso, El Paso, USA
4 shared publications - Paul Hand
Massachusetts Institute of Technology, Department of Mathematics, 77 Massachusetts Avenue E18-369, Cambridge, MA 02139, USA
2 shared publications - Xiangxiong Zhang
Department of Mathematics Purdue University 47907 West Lafayette IN USA
2 shared publications - Adrien Scheuer
Formerly Massachusetts Institute of Technology, Department of Mathematics and Earth Resources Laboratory, 77 Massachusetts Ave, Cambridge, Massachusetts 02139, USA; currently Université Catholique de Louvain, Institute of Information and Communication Technologies, Electronics and Applied Mathematics, 4, Avenue G. Lemaitre, Louvain-la-Neuve B-1348, Belgium.
1 shared publication - Aimé Fournier
Earth Resources Laboratory, Massachusetts Institute of Technology, Cambridge, MA 02139, USA
1 shared publication - Augustin Cosse
Département de Mathématiques et Applications, École Normale Supérieure, Paris, France
1 shared publication
Latest Publications
- Open Access
Superresolution with the zero-phase imaging condition
Wave-based imaging techniques use wavefield data from receivers on the boundary of a domain to produce an image of the underlying structure in the domain of interest. These images are defined by the imaging condition, which maps recorded data to their reflection points in the domain. In this paper, we introduce a nonlinear modification to the standard imaging condition that can produce images with resolutions greater than that ordinarily expected using the standard imaging condition. We show that the phase of the integrand of the imaging condition, in the Fourier domain, has a special significance in some settings that can be exploited to derive a super-resolved modification of the imaging condition. Whereas standard imaging techniques can resolve features of a length scale of $λ$, our technique allows for resolution level $R < λ$, where the super-resolution factor (SRF) is typically $λ/R$. We show that, in the presence of noise, $R \sim σ$.Open AccessSuperresolution with the zero-phase imaging condition
Wave-based imaging techniques use wavefield data from receivers on the boundary of a domain to produce an image of the underlying structure in the domain of interest. These images are defined by the imaging condition, which maps recorded data to their reflection points in the domain. In this paper, we introduce a nonlinear modification to the standard imaging condition that can produce images with resolutions greater than that ordinarily expected using the standard imaging condition. We show that the phase of the integrand of the imaging condition, in the Fourier domain, has a special significance in some settings that can be exploited to derive a super-resolved modification of the imaging condition. Whereas standard imaging techniques can resolve features of a length scale of $λ$, our technique allows for resolution level $R < λ$, where the super-resolution factor (SRF) is typically $λ/R$. We show that, in the presence of noise, $R \sim σ$. - 26 October 2022
- Published by Society for Industrial & Applied Mathematics (SIAM) in Multiscale Modeling & Simulation
- Vol. 20 (4) , 1191-1227
Wide-Band Butterfly Network: Stable and Efficient Inversion Via Multi-Frequency Neural Networks
We introduce an end-to-end deep learning architecture called the wide-band butterfly network (WideBNet) for approximating the inverse scattering map from wide-band scattering data. This a...
- 26 October 2022
- Published by Society for Industrial & Applied Mathematics (SIAM) in Multiscale Modeling & Simulation
- Vol. 20 (4) , 1191-1227
Wide-Band Butterfly Network: Stable and Efficient Inversion Via Multi-Frequency Neural Networks
We introduce an end-to-end deep learning architecture called the wide-band butterfly network (WideBNet) for approximating the inverse scattering map from wide-band scattering data. This a...
- 12 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 34 (7) , 3385-3396
Beyond Correlations: Deep Learning for Seismic Interferometry
Passive seismic interferometry is a vastly generalized blind deconvolution question, where different paths through the Earth correspond to different channels called Green's functions; the sources are completely incoherent and not shared by the channels, and the question is to estimate paths (channels) that are not present in the dataset. SI, turning noise to signal, has numerous applications, from monitoring industrial activities to crustal structure investigation. No standard method of signal processing will solve SI. Instead, domain scientists resort to a simple cross-correlation operation, a.k.a. correlogram, which can retrieve the Green's function directly, but only under restrictive assumptions of ergodicity (energy equipartitioning) of the random process generating the seismic source. However, in practice, correlograms are not equal to the empirical Green's function, because these assumptions are generally far from being satisfied in realistic situations. In the framework of supervised learning, we propose to train deep neural networks (NNs) to overcome two limitations of correlation-based SI: the temporal limitation of passive recordings and the spatial limitation of the random source distribution. Deep NNs are trained to implicitly find the relationship between the empirical Green's functions and the correlograms and then used to extract the correct Green's functions from ambient noise. The input of the network is correlograms (a virtual shot gather), and the desired output is the empirical Green's function (the active shot gather). The NN can often retrieve Green's functions from 5-min passive recordings with acceptable accuracy in our synthetic example. Although an exact estimation of the source locations may not be necessary, a prior knowledge of the source directionality (through a preliminary beamforming step) is helpful when training the NN to mitigate the challenges associated with inhomogeneous source distributions (directional wave fields). In this work, all the numerical examples are based on the retrieval of P-wave reflections in the exploration scale and are conducted on synthetic data. We use a modified ResNet in our numerical experiments.
- 12 May 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks and Learning Systems
- Vol. 34 (7) , 3385-3396
Beyond Correlations: Deep Learning for Seismic Interferometry
Passive seismic interferometry is a vastly generalized blind deconvolution question, where different paths through the Earth correspond to different channels called Green's functions; the sources are completely incoherent and not shared by the channels, and the question is to estimate paths (channels) that are not present in the dataset. SI, turning noise to signal, has numerous applications, from monitoring industrial activities to crustal structure investigation. No standard method of signal processing will solve SI. Instead, domain scientists resort to a simple cross-correlation operation, a.k.a. correlogram, which can retrieve the Green's function directly, but only under restrictive assumptions of ergodicity (energy equipartitioning) of the random process generating the seismic source. However, in practice, correlograms are not equal to the empirical Green's function, because these assumptions are generally far from being satisfied in realistic situations. In the framework of supervised learning, we propose to train deep neural networks (NNs) to overcome two limitations of correlation-based SI: the temporal limitation of passive recordings and the spatial limitation of the random source distribution. Deep NNs are trained to implicitly find the relationship between the empirical Green's functions and the correlograms and then used to extract the correct Green's functions from ambient noise. The input of the network is correlograms (a virtual shot gather), and the desired output is the empirical Green's function (the active shot gather). The NN can often retrieve Green's functions from 5-min passive recordings with acceptable accuracy in our synthetic example. Although an exact estimation of the source locations may not be necessary, a prior knowledge of the source directionality (through a preliminary beamforming step) is helpful when training the NN to mitigate the challenges associated with inhomogeneous source distributions (directional wave fields). In this work, all the numerical examples are based on the retrieval of P-wave reflections in the exploration scale and are conducted on synthetic data. We use a modified ResNet in our numerical experiments.
- 14 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 60 (01962892) , 1-11
Deep Learning for Low-Frequency Extrapolation of Multicomponent Data in Elastic FWI
Full-waveform inversion (FWI) strongly depends on an accurate starting model to succeed. This is particularly true in the elastic regime: The cycle-skipping phenomenon is more severe in elastic FWI compared to acoustic FWI due to the short S-wave wavelength. In this article, we extend our work on extrapolated FWI (EFWI) by proposing to synthesize the low frequencies of multicomponent elastic seismic records and use those “artificial” low frequencies to seed the frequency sweep of elastic FWI. Our solution involves deep learning: we can either train the same convolutional neural network (CNN) on two training datasets, one with vertical components and one with horizontal components of particle velocities, or train with two components together, to extrapolate the low frequencies of elastic data for 2-D elastic FWI. The architecture of this CNN is designed with a large receptive field by dilated convolution. Numerical examples on the Marmousi2 model show that the 2–4 Hz low-frequency data extrapolated from band-limited data above 4 Hz provide good starting models for elastic FWI of P- and S-wave velocities. In addition, we study the generalization ability of the proposed neural network from acoustic to elastic data. For elastic test data, collecting the training dataset by elastic simulation shows better extrapolation accuracy than acoustic simulation, i.e., a smaller generalization gap.
- 14 December 2021
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 60 (01962892) , 1-11
Deep Learning for Low-Frequency Extrapolation of Multicomponent Data in Elastic FWI
Full-waveform inversion (FWI) strongly depends on an accurate starting model to succeed. This is particularly true in the elastic regime: The cycle-skipping phenomenon is more severe in elastic FWI compared to acoustic FWI due to the short S-wave wavelength. In this article, we extend our work on extrapolated FWI (EFWI) by proposing to synthesize the low frequencies of multicomponent elastic seismic records and use those “artificial” low frequencies to seed the frequency sweep of elastic FWI. Our solution involves deep learning: we can either train the same convolutional neural network (CNN) on two training datasets, one with vertical components and one with horizontal components of particle velocities, or train with two components together, to extrapolate the low frequencies of elastic data for 2-D elastic FWI. The architecture of this CNN is designed with a large receptive field by dilated convolution. Numerical examples on the Marmousi2 model show that the 2–4 Hz low-frequency data extrapolated from band-limited data above 4 Hz provide good starting models for elastic FWI of P- and S-wave velocities. In addition, we study the generalization ability of the proposed neural network from acoustic to elastic data. For elastic test data, collecting the training dataset by elastic simulation shows better extrapolation accuracy than acoustic simulation, i.e., a smaller generalization gap.
- Open Access
Wide-band butterfly network: stable and efficient inversion via multi-frequency neural networks
We introduce an end-to-end deep learning architecture called the wide-band butterfly network (WideBNet) for approximating the inverse scattering map from wide-band scattering data. This architecture incorporates tools from computational harmonic analysis, such as the butterfly factorization, and traditional multi-scale methods, such as the Cooley-Tukey FFT algorithm, to drastically reduce the number of trainable parameters to match the inherent complexity of the problem. As a result WideBNet is efficient: it requires fewer training points than off-the-shelf architectures, and has stable training dynamics, thus it can rely on standard weight initialization strategies. The architecture automatically adapts to the dimensions of the data with only a few hyper-parameters that the user must specify. WideBNet is able to produce images that are competitive with optimization-based approaches, but at a fraction of the cost, and we also demonstrate numerically that it learns to super-resolve scatterers in the full aperture scattering setup.
Open AccessWide-band butterfly network: stable and efficient inversion via multi-frequency neural networks
We introduce an end-to-end deep learning architecture called the wide-band butterfly network (WideBNet) for approximating the inverse scattering map from wide-band scattering data. This architecture incorporates tools from computational harmonic analysis, such as the butterfly factorization, and traditional multi-scale methods, such as the Cooley-Tukey FFT algorithm, to drastically reduce the number of trainable parameters to match the inherent complexity of the problem. As a result WideBNet is efficient: it requires fewer training points than off-the-shelf architectures, and has stable training dynamics, thus it can rely on standard weight initialization strategies. The architecture automatically adapts to the dimensions of the data with only a few hyper-parameters that the user must specify. WideBNet is able to produce images that are competitive with optimization-based approaches, but at a fraction of the cost, and we also demonstrate numerically that it learns to super-resolve scatterers in the full aperture scattering setup.
- 2 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 59 (9) , 8034-8039
Lift and Relax for PDE-Constrained Inverse Problems in Seismic Imaging
We present lift and relax for waveform inversion (LRWI), an approach that mitigates the local minima issue in seismic full waveform inversion (FWI) via a combination of two convexification techniques. The first technique (Lift) extends the set of unknown variables to their products, arranged as a moment matrix. This algebraic idea is a celebrated way to replace a hard polynomial optimization problem by a semidefinite programming approximation. Concretely, both the model and the wavefield are lifted from vectors to rank-2 matrices. The second technique (Relax) invites to relax the strict wave-equation constraint--a technique known as wavefield reconstruction inversion (WRI), which introduces wave-equation misfits as a weighted penalty term in the objective function. The relaxed penalty formulation enables balancing the data and wave-equation misfits by tuning a penalty parameter. Together, ``Lift'' and ``Relax'' help reformulate the inverse problem as a set of constraints on a rank-2 moment matrix. Such a lifting strategy permits good data and wave equation fits throughout the inversion process while leaving the numerical rank of the rank-2 moment matrix to be minimized down to one. Moreover, LRWI does not require adjoint wavefield to compute the gradient, which mitigates computational burdens. Numerical examples indicate that starting with a poor initial model, LRWI can conduct successful inversions with a starting frequency that is higher than that required by FWI and WRI.
- 2 December 2020
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Geoscience and Remote Sensing
- Vol. 59 (9) , 8034-8039
Lift and Relax for PDE-Constrained Inverse Problems in Seismic Imaging
We present lift and relax for waveform inversion (LRWI), an approach that mitigates the local minima issue in seismic full waveform inversion (FWI) via a combination of two convexification techniques. The first technique (Lift) extends the set of unknown variables to their products, arranged as a moment matrix. This algebraic idea is a celebrated way to replace a hard polynomial optimization problem by a semidefinite programming approximation. Concretely, both the model and the wavefield are lifted from vectors to rank-2 matrices. The second technique (Relax) invites to relax the strict wave-equation constraint--a technique known as wavefield reconstruction inversion (WRI), which introduces wave-equation misfits as a weighted penalty term in the objective function. The relaxed penalty formulation enables balancing the data and wave-equation misfits by tuning a penalty parameter. Together, ``Lift'' and ``Relax'' help reformulate the inverse problem as a set of constraints on a rank-2 moment matrix. Such a lifting strategy permits good data and wave equation fits throughout the inversion process while leaving the numerical rank of the rank-2 moment matrix to be minimized down to one. Moreover, LRWI does not require adjoint wavefield to compute the gradient, which mitigates computational burdens. Numerical examples indicate that starting with a poor initial model, LRWI can conduct successful inversions with a starting frequency that is higher than that required by FWI and WRI.
- 25 September 2020
- Published by Springer Nature in Foundations of Computational Mathematics
- Vol. 21 (4) , 891-940
Stable Rank-One Matrix Completion is Solved by the Level 2 Lasserre Relaxation
- 25 September 2020
- Published by Springer Nature in Foundations of Computational Mathematics
- Vol. 21 (4) , 891-940
Stable Rank-One Matrix Completion is Solved by the Level 2 Lasserre Relaxation
- Top Cited Papers
- 4 March 2020
- Published by Society of Exploration Geophysicists in Geophysics
- Vol. 85 (3) , R275-R288
Extrapolated full-waveform inversion with deep learning
The lack of low-frequency information and a good initial model can seriously affect the success of full-waveform inversion (FWI), due to the inherent cycle skipping problem. Computational low-frequency extrapolation is in principle the most direct way to address this issue. By considering bandwidth extension as a regression problem in machine learning, we have adopted an architecture of convolutional neural network (CNN) to automatically extrapolate the missing low frequencies. The band-limited recordings are the inputs of the CNN, and, in our numerical experiments, a neural network trained from enough samples can predict a reasonable approximation to the seismograms in the unobserved low-frequency band, in phase and in amplitude. The numerical experiments considered are set up on simulated P-wave data. In extrapolated FWI (EFWI), the low-wavenumber components of the model are determined from the extrapolated low frequencies, before proceeding with a frequency sweep of the band-limited data. The introduced deep-learning method of low-frequency extrapolation shows adequate generalizability for the initialization step of EFWI. Numerical examples show that the neural network trained on several submodels of the Marmousi model is able to predict the low frequencies for the BP 2004 benchmark model. Additionally, the neural network can robustly process seismic data with uncertainties due to the existence of random noise, a poorly known source wavelet, and a different finite-difference scheme in the forward modeling operator. Finally, this approach is not subject to strong assumptions on signals or velocity models of other methods for bandwidth extension and seems to offer a tantalizing solution to the problem of properly initializing FWI.
- 4 March 2020
- Published by Society of Exploration Geophysicists in Geophysics
- Vol. 85 (3) , R275-R288
Top Cited PapersExtrapolated full-waveform inversion with deep learning
The lack of low-frequency information and a good initial model can seriously affect the success of full-waveform inversion (FWI), due to the inherent cycle skipping problem. Computational low-frequency extrapolation is in principle the most direct way to address this issue. By considering bandwidth extension as a regression problem in machine learning, we have adopted an architecture of convolutional neural network (CNN) to automatically extrapolate the missing low frequencies. The band-limited recordings are the inputs of the CNN, and, in our numerical experiments, a neural network trained from enough samples can predict a reasonable approximation to the seismograms in the unobserved low-frequency band, in phase and in amplitude. The numerical experiments considered are set up on simulated P-wave data. In extrapolated FWI (EFWI), the low-wavenumber components of the model are determined from the extrapolated low frequencies, before proceeding with a frequency sweep of the band-limited data. The introduced deep-learning method of low-frequency extrapolation shows adequate generalizability for the initialization step of EFWI. Numerical examples show that the neural network trained on several submodels of the Marmousi model is able to predict the low frequencies for the BP 2004 benchmark model. Additionally, the neural network can robustly process seismic data with uncertainties due to the existence of random noise, a poorly known source wavelet, and a different finite-difference scheme in the forward modeling operator. Finally, this approach is not subject to strong assumptions on signals or velocity models of other methods for bandwidth extension and seems to offer a tantalizing solution to the problem of properly initializing FWI.
- Open Access
Extrapolated full waveform inversion with deep learning
The lack of low frequency information and a good initial model can seriously affect the success of full waveform inversion (FWI), due to the inherent cycle skipping problem. Computational low frequency extrapolation is in principle the most direct way to address this issue. By considering bandwidth extension as a regression problem in machine learning, we propose an architecture of convolutional neural network (CNN) to automatically extrapolate the missing low frequencies without preprocessing and post-processing steps. The bandlimited recordings are the inputs of the CNN and, in our numerical experiments, a neural network trained from enough samples can predict a reasonable approximation to the seismograms in the unobserved low frequency band, both in phase and in amplitude. The numerical experiments considered are set up on simulated P-wave data. In extrapolated FWI (EFWI), the low-wavenumber components of the model are determined from the extrapolated low frequencies, before proceeding with a frequency sweep of the bandlimited data. The proposed deep-learning method of low-frequency extrapolation shows adequate generalizability for the initialization step of EFWI. Numerical examples show that the neural network trained on several submodels of the Marmousi model is able to predict the low frequencies for the BP 2004 benchmark model. Additionally, the neural network can robustly process seismic data with uncertainties due to the existence of noise, poorly-known source wavelet, and different finite-difference scheme in the forward modeling operator. Finally, this approach is not subject to the structural limitations of other methods for bandwidth extension, and seems to offer a tantalizing solution to the problem of properly initializing FWI.
Open AccessExtrapolated full waveform inversion with deep learning
The lack of low frequency information and a good initial model can seriously affect the success of full waveform inversion (FWI), due to the inherent cycle skipping problem. Computational low frequency extrapolation is in principle the most direct way to address this issue. By considering bandwidth extension as a regression problem in machine learning, we propose an architecture of convolutional neural network (CNN) to automatically extrapolate the missing low frequencies without preprocessing and post-processing steps. The bandlimited recordings are the inputs of the CNN and, in our numerical experiments, a neural network trained from enough samples can predict a reasonable approximation to the seismograms in the unobserved low frequency band, both in phase and in amplitude. The numerical experiments considered are set up on simulated P-wave data. In extrapolated FWI (EFWI), the low-wavenumber components of the model are determined from the extrapolated low frequencies, before proceeding with a frequency sweep of the bandlimited data. The proposed deep-learning method of low-frequency extrapolation shows adequate generalizability for the initialization step of EFWI. Numerical examples show that the neural network trained on several submodels of the Marmousi model is able to predict the low frequencies for the BP 2004 benchmark model. Additionally, the neural network can robustly process seismic data with uncertainties due to the existence of noise, poorly-known source wavelet, and different finite-difference scheme in the forward modeling operator. Finally, this approach is not subject to the structural limitations of other methods for bandwidth extension, and seems to offer a tantalizing solution to the problem of properly initializing FWI.
- Open Access
- 1 July 2019
- Published by Society of Exploration Geophysicists in Geophysics
- Vol. 84 (4) , T313-T333
The method of polarized traces for the 3D Helmholtz equation
We have developed a fast solver for the 3D Helmholtz equation, in heterogeneous, constant density, acoustic media, in the high-frequency regime. The solver is based on the method of polarized traces, a layered domain-decomposition method, where the subdomains are connected via transmission conditions prescribed by the discrete Green’s representation formula and artificial reflections are avoided by enforcing nonreflecting boundary conditions between layers. The method of polarized traces allows us to consider only unknowns at the layer interfaces, reducing the overall cost and memory footprint of the solver. We determine that polarizing the wavefields in this manner yields an efficient preconditioner for the reduced system, whose rate of convergence is independent of the problem frequency. The resulting preconditioned system is solved iteratively using generalized minimum residual, where we never assemble the reduced system or preconditioner; rather, we implement them via solving the Helmholtz equation locally within the subdomains. The method is parallelized using Message Passing Interface and coupled with a distributed linear algebra library and pipelining to obtain an empirical on-line runtime [ ], where [ ] is the total number of degrees of freedom, [ ] is the number of subdomains, and [ ] is the number of right-hand sides (RHS). This scaling is favorable for regimes in which the number of sources (distinct RHS) is large, for example, enabling large-scale implementations of frequency-domain full-waveform inversion.
- 1 July 2019
- Published by Society of Exploration Geophysicists in Geophysics
- Vol. 84 (4) , T313-T333
Open AccessThe method of polarized traces for the 3D Helmholtz equation
We have developed a fast solver for the 3D Helmholtz equation, in heterogeneous, constant density, acoustic media, in the high-frequency regime. The solver is based on the method of polarized traces, a layered domain-decomposition method, where the subdomains are connected via transmission conditions prescribed by the discrete Green’s representation formula and artificial reflections are avoided by enforcing nonreflecting boundary conditions between layers. The method of polarized traces allows us to consider only unknowns at the layer interfaces, reducing the overall cost and memory footprint of the solver. We determine that polarizing the wavefields in this manner yields an efficient preconditioner for the reduced system, whose rate of convergence is independent of the problem frequency. The resulting preconditioned system is solved iteratively using generalized minimum residual, where we never assemble the reduced system or preconditioner; rather, we implement them via solving the Helmholtz equation locally within the subdomains. The method is parallelized using Message Passing Interface and coupled with a distributed linear algebra library and pipelining to obtain an empirical on-line runtime [ ], where [ ] is the total number of degrees of freedom, [ ] is the number of subdomains, and [ ] is the number of right-hand sides (RHS). This scaling is favorable for regimes in which the number of sources (distinct RHS) is large, for example, enabling large-scale implementations of frequency-domain full-waveform inversion.