We report research results on the training and use of an artificial neural network for the preconditioning of large linear systems. The aim is to solve a large linear system, which is a discretisation of a Helmholtz differential equation, as quickly as possible using a preconditioned Krylov-like iterative method such as Flexible GMRES. The implementations are written in Python and are divided into two main repositories: one for the Flexible GMRES algorithm, and another for learning neural networks. Currently, a preconditioning matrix is built column by column by solving a set of large sparse block structured least squares problems. This method is relatively easy to implement, but suffers from high computational cost and consequently high en...
Juillet-Septembre 1997In my PhD Thesis ,three constructive algorithms were developed. These algorith...
A classic approach for solving differential equations with neural networks builds upon neural forms,...
In the field of machine learning, deep neural networks have become the inescapablereference for a ve...
The paper proposes a general framework which encompasses the training of neural networks, the adapta...
The structure of a neural network determines to a large extent its cost of training and use, as well...
In many industrials problems, a single evaluation of objective function is expensive in time calcula...
The stochastic gradient method is currently the prevailing technology for training neural networks. ...
Solving differential equations is still a topic of major interest, due to their appearance in many f...
This thesis is a study of practical methods to estimate value functions with feedforward neural netw...
Reinforcement learning describes how an agent can learn to act in an unknown environment in order to...
Neural network models became highly popular during the last decade due to their efficiency in variou...
Cette thèse explore deux idées différentes: — Une méthode améliorée d’entraînement de réseaux de neu...
This thesis analyzes the use of adaptive preconditioned Krylov methods in applications which can be ...
Dans un problème de prédiction à multiples pas discrets, la prédiction à chaque instant peut dépendr...
In the framework of nonlinear process modeling, we propose training algorithms for feedback wavelet ...
Juillet-Septembre 1997In my PhD Thesis ,three constructive algorithms were developed. These algorith...
A classic approach for solving differential equations with neural networks builds upon neural forms,...
In the field of machine learning, deep neural networks have become the inescapablereference for a ve...
The paper proposes a general framework which encompasses the training of neural networks, the adapta...
The structure of a neural network determines to a large extent its cost of training and use, as well...
In many industrials problems, a single evaluation of objective function is expensive in time calcula...
The stochastic gradient method is currently the prevailing technology for training neural networks. ...
Solving differential equations is still a topic of major interest, due to their appearance in many f...
This thesis is a study of practical methods to estimate value functions with feedforward neural netw...
Reinforcement learning describes how an agent can learn to act in an unknown environment in order to...
Neural network models became highly popular during the last decade due to their efficiency in variou...
Cette thèse explore deux idées différentes: — Une méthode améliorée d’entraînement de réseaux de neu...
This thesis analyzes the use of adaptive preconditioned Krylov methods in applications which can be ...
Dans un problème de prédiction à multiples pas discrets, la prédiction à chaque instant peut dépendr...
In the framework of nonlinear process modeling, we propose training algorithms for feedback wavelet ...
Juillet-Septembre 1997In my PhD Thesis ,three constructive algorithms were developed. These algorith...
A classic approach for solving differential equations with neural networks builds upon neural forms,...
In the field of machine learning, deep neural networks have become the inescapablereference for a ve...