-
MathWriting: A Dataset For Handwritten Mathematical Expression Recognition
Authors:
Philippe Gervais,
Asya Fadeeva,
Andrii Maksai
Abstract:
We introduce MathWriting, the largest online handwritten mathematical expression dataset to date. It consists of 230k human-written samples and an additional 400k synthetic ones. MathWriting can also be used for offline HME recognition and is larger than all existing offline HME datasets like IM2LATEX-100K. We introduce a benchmark based on MathWriting data in order to advance research on both onl…
▽ More
We introduce MathWriting, the largest online handwritten mathematical expression dataset to date. It consists of 230k human-written samples and an additional 400k synthetic ones. MathWriting can also be used for offline HME recognition and is larger than all existing offline HME datasets like IM2LATEX-100K. We introduce a benchmark based on MathWriting data in order to advance research on both online and offline HME recognition.
△ Less
Submitted 16 April, 2024;
originally announced April 2024.
-
Conditional integrability and stability for the homogeneous Boltzmann equation with very soft potentials
Authors:
Ricardo J. Alonso,
Pierre Gervais,
Bertrand Lods
Abstract:
We introduce a practical criterion that justifies the propagation and appearance of $L^{p}$-norms for the solutions to the spatially homogeneous Boltzmann equation with very soft potentials without cutoff. Such criterion also provides a new conditional stability result for classical solutions to the equation. All results are quantitative. Our approach is inspired by a recent analogous result for t…
▽ More
We introduce a practical criterion that justifies the propagation and appearance of $L^{p}$-norms for the solutions to the spatially homogeneous Boltzmann equation with very soft potentials without cutoff. Such criterion also provides a new conditional stability result for classical solutions to the equation. All results are quantitative. Our approach is inspired by a recent analogous result for the Landau equation derived in arXiv:2306.15729 and generalises existing conditional results related to higher integrability properties and stability of solutions to the Boltzmann equation with very soft potentials.
△ Less
Submitted 22 March, 2024;
originally announced March 2024.
-
Hydrodynamic limits for kinetic equations preserving mass, momentum and energy: a spectral and unified approach in the presence of a spectral gap
Authors:
Pierre Gervais,
Bertrand Lods
Abstract:
Triggered by the fact that, in the hydrodynamic limit, several different kinetic equations of physical interest all lead to the same Navier-Stokes-Fourier system, we develop in the paper an abstract framework which allows to explain this phenomenon. The method we develop can be seen as a significant improvement of known approaches for which we fully exploit some structural assumptions on the linea…
▽ More
Triggered by the fact that, in the hydrodynamic limit, several different kinetic equations of physical interest all lead to the same Navier-Stokes-Fourier system, we develop in the paper an abstract framework which allows to explain this phenomenon. The method we develop can be seen as a significant improvement of known approaches for which we fully exploit some structural assumptions on the linear and nonlinear collision operators as well as a good knowledge of the Cauchy theory for the limiting equation. {In particular, we fully exploit the fact that the collision operator is preserving both momentum and kinetic energy.} We adopt a perturbative framework in a Hilbert space setting and first develop a general and fine spectral analysis of the linearized operator and its associated semigroup. Then, we introduce a splitting adapted to the various regimes (kinetic, acoustic, hydrodynamic) present in the kinetic equation which allows, by a fixed point argument, to construct a solution to the kinetic equation and prove the convergence towards suitable solutions to the Navier-Stokes-Fourier system. Our approach is robust enough to treat, in the same formalism, the case of the Boltzmann equation with hard and moderately soft potentials, with and without cut-off assumptions, as well as the Landau equation for hard and moderately soft potentials in presence of a spectral gap. New well-posedness and strong convergence results are obtained within this framework. In particular, for initial data with algebraic decay with respect to the velocity variable, our approach provides the first result concerning the strong Navier-Stokes limit from Boltzmann equation without Grad cut-off assumption or Landau equation. The method developed in the paper is also robust enough to apply, at least at the linear level, to quantum kinetic equations for Fermi-Dirac or Bose-Einstein particles.
△ Less
Submitted 10 April, 2024; v1 submitted 23 April, 2023;
originally announced April 2023.
-
Non-cutoff Boltzmann equation with soft potentials in the whole space
Authors:
Kleber Carrapatoso,
Pierre Gervais
Abstract:
We prove the existence and uniqueness of global solutions to the Boltzmann equation with non-cutoff soft potentials in the whole space when the initial data is a small perturbation of a Maxwellian with polynomial decay in velocity. Our method is based in the decomposition of the desired solution into two parts: one with polynomial decay in velocity satisfying the Boltzmann equation with only a dis…
▽ More
We prove the existence and uniqueness of global solutions to the Boltzmann equation with non-cutoff soft potentials in the whole space when the initial data is a small perturbation of a Maxwellian with polynomial decay in velocity. Our method is based in the decomposition of the desired solution into two parts: one with polynomial decay in velocity satisfying the Boltzmann equation with only a dissipative part of the linearized operator ; the other with Gaussian decay in velocity verifying the Boltzmann equation with a coupling term.
△ Less
Submitted 28 June, 2023; v1 submitted 8 December, 2022;
originally announced December 2022.
-
On the convergence from Boltzmann to Navier-Stokes-Fourier for general initial data
Authors:
Pierre Gervais
Abstract:
In this work, we prove the convergence of strong solutions of the Boltzman equation, for initial data having polynomial decay in the velocity variable, towards those of the incompressible Navier-Stokes-Fourier system. We show in particular that the solutions of the rescaled Boltzmann equation do not blow up before their hydrodynamic limit does. This is made possible by adapting a strategy introduc…
▽ More
In this work, we prove the convergence of strong solutions of the Boltzman equation, for initial data having polynomial decay in the velocity variable, towards those of the incompressible Navier-Stokes-Fourier system. We show in particular that the solutions of the rescaled Boltzmann equation do not blow up before their hydrodynamic limit does. This is made possible by adapting a strategy introduced by M. Briant, S. Merino and C. Mouhot of writing the solution to the Boltzmann equation as the sum a part with polynomial decay and a second one with Gaussian decay. The Gaussian part is treated with an approach reminiscent of the one used by I. Gallagher and I. Tristani.
△ Less
Submitted 8 January, 2022;
originally announced January 2022.
-
Spectral study of the linearized Boltzmann operator in L^2 spaces with polynomial and Gaussian weights
Authors:
Pierre Gervais
Abstract:
The aim of this paper is to extend to the spaces L^2(R^d , (1+|v|)^2k dv) the spectral study led in L^2(R^d , exp(|v|^2/2)dv) by R. Ellis and M. Pinsky on the space inhomogeneous linearized Boltzmann operator for hard spheres. More precisely, we look at the Fourier transform in the space variable of the inhomogeneous operator and consider the dual Fourier variable as a fixed parameter. We then per…
▽ More
The aim of this paper is to extend to the spaces L^2(R^d , (1+|v|)^2k dv) the spectral study led in L^2(R^d , exp(|v|^2/2)dv) by R. Ellis and M. Pinsky on the space inhomogeneous linearized Boltzmann operator for hard spheres. More precisely, we look at the Fourier transform in the space variable of the inhomogeneous operator and consider the dual Fourier variable as a fixed parameter. We then perform a precise study of this operator for small frequencies (by seeing it as a perturbation of the homogeneous one) and also for large frequencies from spectral and semigroup point of views. Our approach is based on perturbation theory for linear operators as well as enlargement arguments from M.P. Gualdani, S. Mischler and C. Mouhot.
△ Less
Submitted 20 October, 2020;
originally announced October 2020.
-
The DIDI dataset: Digital Ink Diagram data
Authors:
Philippe Gervais,
Thomas Deselaers,
Emre Aksan,
Otmar Hilliges
Abstract:
We are releasing a dataset of diagram drawings with dynamic drawing information. The dataset aims to foster research in interactive graphical symbolic understanding. The dataset was obtained using a prompted data collection effort.
We are releasing a dataset of diagram drawings with dynamic drawing information. The dataset aims to foster research in interactive graphical symbolic understanding. The dataset was obtained using a prompted data collection effort.
△ Less
Submitted 24 February, 2020; v1 submitted 20 February, 2020;
originally announced February 2020.
-
Fast Multi-language LSTM-based Online Handwriting Recognition
Authors:
Victor Carbune,
Pedro Gonnet,
Thomas Deselaers,
Henry A. Rowley,
Alexander Daryin,
Marcos Calvo,
Li-Lun Wang,
Daniel Keysers,
Sandro Feuz,
Philippe Gervais
Abstract:
We describe an online handwriting system that is able to support 102 languages using a deep neural network architecture. This new system has completely replaced our previous Segment-and-Decode-based system and reduced the error rate by 20%-40% relative for most languages. Further, we report new state-of-the-art results on IAM-OnDB for both the open and closed dataset setting. The system combines m…
▽ More
We describe an online handwriting system that is able to support 102 languages using a deep neural network architecture. This new system has completely replaced our previous Segment-and-Decode-based system and reduced the error rate by 20%-40% relative for most languages. Further, we report new state-of-the-art results on IAM-OnDB for both the open and closed dataset setting. The system combines methods from sequence recognition with a new input encoding using Bézier curves. This leads to up to 10x faster recognition times compared to our previous system. Through a series of experiments we determine the optimal configuration of our models and report the results of our setup on a number of additional public datasets.
△ Less
Submitted 24 January, 2020; v1 submitted 22 February, 2019;
originally announced February 2019.
-
Machine Learning for Neuroimaging with Scikit-Learn
Authors:
Alexandre Abraham,
Fabian Pedregosa,
Michael Eickenberg,
Philippe Gervais,
Andreas Muller,
Jean Kossaifi,
Alexandre Gramfort,
Bertrand Thirion,
Gäel Varoquaux
Abstract:
Statistical machine learning methods are increasingly used for neuroimaging data analysis. Their main virtue is their ability to model high-dimensional datasets, e.g. multivariate analysis of activation images or resting-state time series. Supervised learning is typically used in decoding or encoding settings to relate brain images to behavioral or clinical observations, while unsupervised learnin…
▽ More
Statistical machine learning methods are increasingly used for neuroimaging data analysis. Their main virtue is their ability to model high-dimensional datasets, e.g. multivariate analysis of activation images or resting-state time series. Supervised learning is typically used in decoding or encoding settings to relate brain images to behavioral or clinical observations, while unsupervised learning can uncover hidden structures in sets of images (e.g. resting state functional MRI) or find sub-populations in large cohorts. By considering different functional neuroimaging applications, we illustrate how scikit-learn, a Python machine learning library, can be used to perform some key analysis steps. Scikit-learn contains a very large set of statistical learning algorithms, both supervised and unsupervised, and its application to neuroimaging data provides a versatile tool to study the brain.
△ Less
Submitted 12 December, 2014;
originally announced December 2014.
-
Lagrangian Vorticity and Velocity Measurements in Turbulent Jets
Authors:
C. Poulain,
N. Mazellier,
P. Gervais,
Y. Gagne,
C. Baudet
Abstract:
We report an experimental investigation of various statistical properties of the spatial Fourier modes of the vorticity field in turbulent jets for a large range of Reynolds numbers ($530 \leq R_λ \leq 6100$). The continuous time evolution of a spatial Fourier mode of the vorticity distribution, characterized by a well defined wave-vector, is obtained from acoustic scattering measurements. The s…
▽ More
We report an experimental investigation of various statistical properties of the spatial Fourier modes of the vorticity field in turbulent jets for a large range of Reynolds numbers ($530 \leq R_λ \leq 6100$). The continuous time evolution of a spatial Fourier mode of the vorticity distribution, characterized by a well defined wave-vector, is obtained from acoustic scattering measurements. The spatial enstrophy spectrum, as a function of the spatial wave-vector, is determined by scanning the incoming sound frequencies. Time-frequency analysis of the turbulent vorticity fluctuations is also performed for different length scales of the flows. Vorticity time-correlations show that the characteristic time of a Fourier mode behaves as the swee** time. Finally, we report preliminary Lagrangian velocity measurements obtained using acoustic scattering by soap bubbles inflated with helium. Gathering a very large number ($\geq 13~10^{6}$) of passages of isolated bubbles in the scattering volume, one is able to compute the Lagrangian velocity PDF and velocity spectrum. Despite the spatial filtering due to the finite size of the bubble, the latter exhibits a power law with the -2 exponent predicted by the Komogorov theory, over one decade of frequencies.
△ Less
Submitted 31 May, 2003;
originally announced June 2003.
