Big ideas for small data

Le prochain atelier du GdR, intitulé "Big ideas for small data" et organisé par Merlin Keller, aura lieu le 13 mai à l'IHP.

Atelier co-organisé avec le groupe "Fiabilité et incertitude" de la SFdS.

Programme et inscription :  ici !

Summer (autumn) schools

Un petit point sur quelques écoles d'été (et d'automne...) 2019.

Sur des thèmes très MASCOT :

• ETICS 2019 : September 23-27, Fréjus (France)
  École Thématique sur les Incertitudes en Calcul Scientifique
  [http://www.gdr-mascotnum.fr/etics.html]
• MNMUQ 2019 : September 2-6, Porquerolles Island (France)
  Modeling and Numerical Methods for Uncertainty Quantification
  [https://www.sigma-clermont.fr/en/mnmuq2019]
• GPSS 2019 : September 9-12, Sheffield (UK)
  Gaussian Process and Uncertainty Quantification Summer School
  [http://gpss.cc]

Un peu moins directement "MASCOT" (?) :

• NASS 2019 :  May 13-17, Paris (France)
  CEA-EDF-INRIA Numerical Analysis Summer School
  "A numerical introduction to optimal transport"
  [http://www-hpc.cea.fr/SummerSchools/SummerSchools2019-NASS.htm]
• SPARISTY4PSL : June 24-27, Paris (France)
  Sparsity for Physics, Signal and Learning
  [https://sparsity4psl.github.io]

N'hésitez pas à indiquer d'autres écoles intéressantes, en commentaire de ce post ou en complétant la page http://www.gdr-mascotnum.fr/schools.html.

Appel à contribution au séminaire « Appréhender la grande dimension »

Les « Rencontres chercheurs et ingénieurs » de Phimeca visent à stimuler la créativité des chercheurs et des ingénieurs pour la résolution de problèmes mettant en jeu la simulation numérique et l’analyse de données, et à encourager les transferts de savoirs et de technologies entre communautés scientifiques.

Elles sont gratuites et ouvertes à tous.

Le jeudi 6 juin 2019 aura lieu à l’institut Henri Poincaré à Paris une Rencontre chercheurs et ingénieurs intitulée « Appréhender la grande dimension ».

Consultez le site seminaire.phimeca.com pour plus de détails.

Vous pouvez vous inscrire ou proposer un exposé en écrivant à seminaire@phimeca.com.

Openturns, an open source initiative for the Treatment of Uncertainties, Risks'N Statistics: version 1.12 released!

OpenTURNS is a scientific C++ and Python library including an internal data model and algorithms dedicated to the treatment of uncertainties. The main goal of this library is giving to specific applications all the functionalities needed to treat uncertainties in studies. Targeted users are all engineers who want to introduce the probabilistic dimension in their so far deterministic studies.

OpenTURNS is distributed under the Lesser General Public License. Please see the LICENSE and the COPYING* files for details of the license of each components.

Up-to-date information can be found at http://www.openturns.org.

Release notes

The major changes of version 1.12 are:

• FieldFunction and co knows its input/output meshes before evaluation.
• SobolSequence has been extended from maximum dimension 40 to 1111
• Parametrized statistical tests by first kind risk instead of 1-risk
• FittingTest now compute a correct p-value in Kolmogorov tests even if the parameters are estimated from the tested sample
• Completed documentation migration with process content

Please see the ChangeLog file for a more exhaustive summary of bug fixes and new features for the current release.

Backwards Compatibility

The developers strive their best to preserve backwards compatibility between releases, but this is not always possible. Where backwards compatibility is known to be broken, it is clearly marked as an incompatibility in the ChangeLog file.

Installation

Please see the http://openturns.github.io/openturns/latest/install.html for instructions on installing OpenTURNS on various platforms from binaries or sources.

Documentation

The http://openturns.github.io/openturns/latest/contents.html page contains the most recent set of updated documentation for this release.

Contributing

See http://openturns.github.io/openturns/latest/developer_guide/developer_guide.html for instructions.

-- The OpenTURNS team

25ème Semaine d'Étude Mathématiques-Entreprise (SEME)

La 25ème SEME (Semaine d'Etude Mathématiques-Entreprise) aura lieu à Orsay du 14 au 18 janvier 2019 :

https://www.math.u-psud.fr/seme2019/

Les inscriptions sont ouvertes, et le resteront tant qu'il restera des places disponibles (une trentaine) et jusque début décembre au plus tard.

Les SEME sont organisées par l'agence maths-entreprises sur le modèle "European Study Group in Industry" : des groupes de 5 à 6 doctorants sont invités à travailler pendant une semaine sur des problèmes exploratoires concrets proposés par des industriels. Les doctorants devront proposer des pistes de réponses innovantes aux problèmes posés, et les présenter en public dès la fin de la semaine d'étude.

Les sujets confirmés pour cette SEME à Orsay sont proposés par: SNCF, Darkmatter, Seenergi, Stormancer, Aéroports de Paris et Dataswati.

Il n'y a pas de frais d'inscription, l'organisation de la SEME prend en charge l'hébergement et les déjeuners du lundi 14 au vendredi 18 janvier, ainsi que le dîner du mardi 15 janvier. La participation à l'intégralité de la semaine (du lundi  au vendredi après-midi) est obligatoire. Le transport et les autres dîners sont à faire prendre en charge par le laboratoire d'origine.

MASCOT-NUM 2019: call for PhD student communications

Le bureau du GdR vous informe :

Dear colleagues,

The 15th venue of the MASCOT-NUM annual conference will be held in Rueil-Malmaison, near Paris, from 18th to 20th March, 2019.

Website: http://www.mascotnum2019.com/

The first day of the annual Mascot-Num workshop is devoted to PhD students communications on topics related to the Mascot-Num network activities, i.e. sensitivity analysis, uncertainty quantification in simulation, design and modelling of computer experiments, model calibration and validation, structural reliability, optimization under uncertainty, data-driven modelling methods, etc.

The second day will be dedicated to the 60th anniversary of Luc Pronzato. A serie of presentation centered around in Luc’s favorite research topics is scheduled.

The last day will be devoted to talks of invited speakers on topics linked to the GDR MASCOT NUM activities.

PhD students are invited to submit a two-page abstract using the LaTeX abstract template in attachment. Seven students will be selected by the Mascot-Num Scientific Committee for oral presentations on March 18th, 2019. The other accepted abstracts will be considered for the poster session to be held on the same day.

The PDF file shall be sent by December 14th, 2018 to

gdrmascotnum@gmail.com.

Please specify in your email if you want your abstract to be considered for an oral presentation or a poster. Authors will be informed whether their abstracts have been accepted in oral or poster by 28 January.

A prize of 1000€ will be awarded for the best oral presentation by the Mascot-Num scientific committee. A prize of 500€ will be awarded for the best poster by the Mascot-Num scientific committee.

This event is supported and organized in the framework of the IFP Energies nouvelles and GDR Mascot-Num.

Regards

The GdR MASCOT-NUM board

UQLab - The Uncertainty Quantification Framework - v.1.1 now available

The new version of UQLab (V1.1) was released on July 5th, 2018.

UQLab is a general purpose Uncertainty Quantification software developed at ETH Zurich (Switzerland) made of interacting modules that solve all types of uncertainty quantification problems in engineering, economics and applied sciences: uncertainty propagation through Monte Carlo simulation (e.g. using surrogate models such as Gaussian processes and polynomial chaos expansions), global sensitivity analysis, rare event estimations, etc.

Three brand new modules are released in V1.1:

• UQLink allows for a seamless connection of third-party software to UQLab using universal "wrapping" of external codes through templates and a mark-up system.
• Support vector machines for regression (SVR) and classification (SVC) allow the analyst to build predictive models from data. In the context of uncertainty quantification, SVR can be used as surrogate models of complex simulators using designs of computer experiments. SVC can be used in the context of rare event estimation.

Visit our website www.uqlab.com. Register, download, enjoy!

---

The UQLab Development Team

ETH Zürich | Chair of Risk, Safety & Uncertainty Quantification

Prof. Bruno Sudret