Solving an inverse problem in physical sciences can involve exploring all possible models. We propose here to train a neural network to reproduce physical simulations. This network is then used to sample the set of models explaining the observed data. We use this approach to fit geological models of upper mantle called marble cakes on tomographic data. We show that the generated images are statistically close to the theoretical models and we get the result of the inversion as a distribution.-La résolution d'un problème inverse en sciences physiques peut passer par l'exploration de l'ensemble des modèles possibles. Nous proposons ici d'entraîner un réseau de neurones à reproduire des simulations physiques. Ce réseau est ensuite utilisé pour ...
Casting a geophysical inverse problem into a Bayesian setting is often discouraged by the computatio...
When solving inverse problems in geophysical imaging, deep generative models (DGMs) may be used to e...
The paper considers the problem of performing a task defined on a model parameter that is only obser...
Solving an inverse problem in physical sciences can involve exploring all possible models. We propos...
We implement a machine-learning inversion approach to infer petrophysical rock properties from pre-s...
In recent years with the technology developments, airborne geophysical methods (gravity, magnetic, a...
The solution to a variety of engineering problems entails the simulation of a physical system. The m...
Probabilistic inversion within a multiple‐point statistics framework is often computationally prohib...
Tomographic imaging is in general an ill-posed inverse problem. Typically, a single regularized imag...
This paper presents a deep learning algorithm for tomographic reconstruction (GANrec). The algorithm...
Notre compréhension de la dynamique de la Terre repose sur l'imagerie de sa structure interne par la...
Efficient and high-fidelity prior sampling and inversion for complex geological media is still a lar...
This paper introduces a novel approach to solve inverse problems by leveraging deep learning techniq...
Inverse Problems in medical imaging and computer vision are traditionally solved using purely model-...
Gravity surveys in regional geophysical research can be used to estimate the depth of the sediment-b...
Casting a geophysical inverse problem into a Bayesian setting is often discouraged by the computatio...
When solving inverse problems in geophysical imaging, deep generative models (DGMs) may be used to e...
The paper considers the problem of performing a task defined on a model parameter that is only obser...
Solving an inverse problem in physical sciences can involve exploring all possible models. We propos...
We implement a machine-learning inversion approach to infer petrophysical rock properties from pre-s...
In recent years with the technology developments, airborne geophysical methods (gravity, magnetic, a...
The solution to a variety of engineering problems entails the simulation of a physical system. The m...
Probabilistic inversion within a multiple‐point statistics framework is often computationally prohib...
Tomographic imaging is in general an ill-posed inverse problem. Typically, a single regularized imag...
This paper presents a deep learning algorithm for tomographic reconstruction (GANrec). The algorithm...
Notre compréhension de la dynamique de la Terre repose sur l'imagerie de sa structure interne par la...
Efficient and high-fidelity prior sampling and inversion for complex geological media is still a lar...
This paper introduces a novel approach to solve inverse problems by leveraging deep learning techniq...
Inverse Problems in medical imaging and computer vision are traditionally solved using purely model-...
Gravity surveys in regional geophysical research can be used to estimate the depth of the sediment-b...
Casting a geophysical inverse problem into a Bayesian setting is often discouraged by the computatio...
When solving inverse problems in geophysical imaging, deep generative models (DGMs) may be used to e...
The paper considers the problem of performing a task defined on a model parameter that is only obser...