Êtes-vous un étudiant de l'EPFL à la recherche d'un projet de semestre?

Travaillez avec nous sur des projets en science des données et en visualisation, et déployez votre projet sous forme d'application sur GraphSearch.

Découvrez-en plus sur les applications Graph.

Graph
Search

fr|en

Se Connecter

Concept

Laplacien discret

Résumé

En mathématiques, le laplacien discret est une analogie du laplacien continu adaptée au cas de problèmes discret (graphes, par exemple). Il est notamment employé en analyse numérique, par exemple dans le cadre de la résolution de l'équation de la chaleur par la méthode des différences finies, ou en pour la détection de contours. Définition en dimension 2 Soit une fonction réelle f de deux variables réelles et x et y. On définit le laplacien discret de f comme la somme des dérivées secondes discrètes selon x et selon y, soit : \begin{align} \Delta_{\text{discret}}f &= \frac{f(x+h,y)+f(x-h,y)-2f(x,y)}{h^2} + \frac{f(x,y+h)+f(x,y-h)-2f(x,y)}{h^2} \ & = \frac{f(x+h,y)+f(x-h,y) + f(x,y+h)+f(x,y-h)-4 f(x,y)}{h^2} \end{align} L'exemple précédent est décrit dans une grille régulière cartésienne de dimension d=2 (plan). En chaque point, le calcul du lap

Source officielle

https://fr.wikipedia.org/wiki/Laplacien_discret

À propos de ce résultat

Cette page est générée automatiquement et peut contenir des informations qui ne sont pas correctes, complètes, à jour ou pertinentes par rapport à votre recherche. Il en va de même pour toutes les autres pages de ce site. Veillez à vérifier les informations auprès des sources officielles de l'EPFL.

Publications associées (17)

Publications associées

Chargement

Personnes associées

Chargement

Unités associées

Chargement

Concepts associés (6)

Concepts associés

Chargement

Cours associés (5)

Cours associés

Chargement

Séances de cours associées (16)

Séances de cours associées

Chargement

Graph Laplacians for Rotation Equivariant Convolutional Neural Networks

Martino Milani

A fundamental problem in signal processing is to design computationally efficient algorithms to filter signals. In many applications, the signals to filter lie on a sphere. Meaningful examples of data of this kind are weather data on the Earth, or images of the sky. It is then important to design filtering algorithms that are computationally efficient and capable of exploiting the rotational symmetry of the problem. In these applications, given a continuous signal

f: \mathbb S^2 \rightarrow \mathbb R

on a 2-sphere

\mathbb S^2 \subset \mathbb R^3

, we can only know the vector of its sampled values

\mathbf f \in \mathbb R^N:\ (\mathbf f)_i = f(\mathbf x_i)

in a finite set of points

\mathcal P \subset \mathbb S^2,\quad \mathcal P = \{\mathbf x_i\}_{i=0}^{n-1}

where our sensors are located. Perraudin et al. in \cite{DeepSphere} construct a sparse graph

G

on the vertex set

\mathcal P

and then use a polynomial of the corresponding graph Laplacian matrix

\mathbf L \in \mathbb R^{n\times n}

to perform a computationally efficient -

\mathcal O (n)

- filtering of the sampled signal

\mathbf f

. In order to study how well this algorithm respects the symmetry of the problem - i.e it is equivariant to the rotation group SO(3) - it is important to guarantee that the spectrum of

\mathbf L

and spectrum of the Laplace-Beltrami operator

\Delta_\mathbb S^2

are somewhat ``close''. We study the spectral properties of such graph Laplacian matrix in the special case of \cite{DeepSphere} where the sampling

\mathcal P

is the so called HEALPix sampling (acronym for \textbf Hierarchical \textbf Equal \textbf Area iso\textbf Latitude \textbf {Pix}elization) and we show a way to build a graph

G'

such that the corresponding graph Laplacian matrix

\mathbf L'

shows better spectral properties than the one presented in \cite{DeepSphere}. We investigate other different methods of building the matrix

\mathbf L

better suited to non uniform sampling measures. In particular, we studied the Finite Element Method approximation of the Laplace-Beltrami operator on the sphere, and how FEM filtering relates to graph filtering, showing the importance of non symmetric discrete Laplacians when it comes to non uniform sampling measures. We finish by showing how the graph Laplacian

\mathbf L'

proposed in this work improved the performances of DeepSphere in a well known classification task using different sampling schemes of the sphere, and by comparing the different Discrete Laplacians introduced in this work.

2019

Chargement

Feynman-Kac representation for the parabolic Anderson model driven by fractional noise

Kamran Kalbasi

We consider the parabolic Anderson model driven by fractional noise: partial derivative/partial derivative t u(t,x) = k Delta u(t,x) + u(t,x)partial derivative/partial derivative t W(t,x) x is an element of Z(d), t >= 0, where k > 0 is a diffusion constant, Delta is the discrete Laplacian defined by Delta f (x) = 1/2d Sigma vertical bar y-x vertical bar=1 (f(Y) - f (0)), and {W(t,x) ; t > 0} x is an element of Z(d) is a family of independent fractional Brownian motions with Hurst parameter H is an element of (0,1), indexed by Z(d). We make sense of this equation via a Stratonovich integration obtained by approximating the fractional Brownian motions with a family of Gaussian processes possessing absolutely continuous sample paths. We prove that the Feynman-Kac representation u(t, x) = E-infinity [u(0) (X(t)) exp integral(t)(0) W(ds,X(t-s))], (1) is a mild solution to this problem. Here u(0) (y) is the initial value at site y is an element of Z(d), {X(t); t >= 0} is a simple random walk with jump rate f, started at x E Zd and independent of the family {W(t, x) ; t >= 0}(x is an element of Zd) and E-infinity is expectation withrespect to this random walk. We give a unified argument that works for any Hurst parameter H is an element of (0,1). (C) 2015 Elsevier Inc. All rights reserved.

Academic Press Inc Elsevier Science2015

Chargement

We study the one-dimensional discrete Schrödinger operator with the skew-shift potential

2\lambda\cos\left(2\pi \left(\binom{j}{2} \omega+jy+x\right)\right)

. This potential is long conjectured to behave like a random one, i.e., it is expected to produce Anderson localization for arbitrarily small coupling constants

\lambda>0

. In this paper, we introduce a novel perturbative approach for studying the zero-energy Lyapunov exponent

L(\lambda)

at small

\lambda

. Our main results establish that, to second order in perturbation theory, a natural upper bound on

L(\lambda)

is fully consistent with

L(\lambda)

being positive and satisfying the usual Figotin-Pastur type asymptotics

L(\lambda)\sim C\lambda^2

\lambda\to 0

. The analogous quantity behaves completely differently in the Almost-Mathieu model, whose zero-energy Lyapunov exponent vanishes for $\lambda

2018

Chargement

Graph Laplacians for Rotation Equivariant Convolutional Neural Networks

Martino Milani

f: \mathbb S^2 \rightarrow \mathbb R

on a 2-sphere

\mathbb S^2 \subset \mathbb R^3

, we can only know the vector of its sampled values

\mathbf f \in \mathbb R^N:\ (\mathbf f)_i = f(\mathbf x_i)

in a finite set of points

\mathcal P \subset \mathbb S^2,\quad \mathcal P = \{\mathbf x_i\}_{i=0}^{n-1}

where our sensors are located. Perraudin et al. in \cite{DeepSphere} construct a sparse graph

G

on the vertex set

\mathcal P

and then use a polynomial of the corresponding graph Laplacian matrix

\mathbf L \in \mathbb R^{n\times n}

to perform a computationally efficient -

\mathcal O (n)

- filtering of the sampled signal

\mathbf f

. In order to study how well this algorithm respects the symmetry of the problem - i.e it is equivariant to the rotation group SO(3) - it is important to guarantee that the spectrum of

\mathbf L

and spectrum of the Laplace-Beltrami operator

\Delta_\mathbb S^2

are somewhat ``close''. We study the spectral properties of such graph Laplacian matrix in the special case of \cite{DeepSphere} where the sampling

\mathcal P

is the so called HEALPix sampling (acronym for \textbf Hierarchical \textbf Equal \textbf Area iso\textbf Latitude \textbf {Pix}elization) and we show a way to build a graph

G'

such that the corresponding graph Laplacian matrix

\mathbf L'

shows better spectral properties than the one presented in \cite{DeepSphere}. We investigate other different methods of building the matrix

\mathbf L

\mathbf L'

2019

We study the one-dimensional discrete Schrödinger operator with the skew-shift potential

2\lambda\cos\left(2\pi \left(\binom{j}{2} \omega+jy+x\right)\right)

. This potential is long conjectured to behave like a random one, i.e., it is expected to produce Anderson localization for arbitrarily small coupling constants

\lambda>0

. In this paper, we introduce a novel perturbative approach for studying the zero-energy Lyapunov exponent

L(\lambda)

at small

\lambda

. Our main results establish that, to second order in perturbation theory, a natural upper bound on

L(\lambda)

is fully consistent with

L(\lambda)

being positive and satisfying the usual Figotin-Pastur type asymptotics

L(\lambda)\sim C\lambda^2

\lambda\to 0

. The analogous quantity behaves completely differently in the Almost-Mathieu model, whose zero-energy Lyapunov exponent vanishes for $\lambda

2018