L'IRIF est une unité mixte de recherche (UMR 8243) entre le CNRS et l'université Paris-Diderot, qui héberge deux équipes-projets INRIA.

Les recherches menées à l'IRIF reposent sur l’étude et la compréhension des fondements de toute l’informatique, afin d’apporter des solutions innovantes aux défis actuels et futurs des sciences numériques.

L'IRIF regroupe près de deux cents personnes. Six de ses membres ont été lauréats de l'European Research Council (ERC), trois sont membres de l'Institut Universitaire de France (IUF) et deux sont membres de l'Academia Europæa.

Kamil Khadiev

11.6.2019
Kamil Khadiev from Kazan University, an expert in complexity and quantum computing, is visiting IRIF from June 10th to July 9th.

CIMPA

23.1.2019
A CIMPA school on Graphs, Algorithms and Randomness is co-organized by Reza Naserasr from IRIF at Tabriz University, 15-22 June 2019. Three colleagues from IRIF, Pierre Fraigniaud, Michel Habib and Frédéric Magniez, are among the five lecturers from France.

Conference CAV'19

30.5.2019
Four papers coauthored by IRIF members will be presented at the prestigious conference CAV'19 in New York this summer. Topics include verifying weakly-consistent distributed databases, testing cache coherence protocols, and testing concurrent objects.

perso-mikael-rabie.jpg

25.5.2019
Mikaël Rabie (Postdoc at IRIF) will present at ICALP'19 a new problem, about distributed reconfiguration of maximal independent sets, providing an optimal algorithm to produce a reconfiguration schedule in the LOCAL model. This work, receiving Best Paper award in Track C, is a joint work with Keren Censor-Hillel from the Technion (Haifa, Israël).

(Ces actualités sont présentées selon un classement mêlant priorité et aléatoire.)

Preuves, programmes et systèmes
Lundi 17 juin 2019, 11 heures, Salle 3052
Pierre-Malo Deniélou (Google) From MapReduce to Apache Beam: A Journey in Abstraction

(This is a joint seminar between the CompSys, PPS, and Verification seminar series.)

Processing large amounts of data used to be an affair of specialists: specialized hardware, specialized software, specialized programming model, specialized engineers. MapReduce was the first widely adopted high-level API for large-scale data processing. It helped democratize big data processing by providing a clear abstraction that was supported by several efficient systems. In this talk, I will present how the programming APIs (and underlying systems) for large-scale data processing evolved in the past 20 years, both within Google and in the open source world. I will start from MapReduce and Millwheel and finish with Apache Beam and where we're headed next.

Vérification
Lundi 17 juin 2019, 11 heures, Salle 3052
Pierre-Malo Deniélou (Google) From MapReduce to Apache Beam: A Journey in Abstraction

Processing large amounts of data used to be an affair of specialists: specialized hardware, specialized software, specialized programming model, specialized engineers. MapReduce was the first widely adopted high-level API for large-scale data processing. It helped democratize big data processing by providing a clear abstraction that was supported by several efficient systems.

In this talk, I will present how the programming APIs (and underlying systems) for large-scale data processing evolved in the past 20 years, both within Google and in the open source world. I will start from MapReduce and Millwheel and finish with Apache Beam and where we're headed next.

(*) This will be a joint session of the Systèmes Complexes, Vérification and PPS seminars

Soutenances d'habilitations
Mardi 18 juin 2019, 10 heures, Salle des Thèses, Halle aux Farines
Yves Guiraud Méthodes de réécriture en algèbre supérieure

Preuves, programmes et systèmes
Mercredi 19 juin 2019, 14 heures 30, Salle 3052
Ugo Dal Lago (INRIA and Univ. Bologna) The Geometry of Bayesian Programming

We give a geometry of interaction model for a typed lambda-calculus endowed with operators for sampling from a continuous uniform distribution and soft conditioning, namely a paradigmatic calculus for higher-order Bayesian programming. The model is based on the category of measurable spaces and partial measurable functions, and is proved adequate with respect to both a distribution-based and a sampling based operational semantics.

Joint work with Naohiko Hoshino

Combinatoire énumérative et analytique
Jeudi 20 juin 2019, 11 heures 45, Salle 1007
Vincent Jugé (Université Paris-Est Marne-la-Vallée) Non encore annoncé.

Soutenances de thèses
Jeudi 20 juin 2019, 14 heures 30, Salle 0011
Raphaëlle Crubillé (IRIF) Behavioural distances for higher-order probabilistic programs

The manuscript is available at: http://research.crubille.lautre.net/main.pdf

Catégories supérieures, polygraphes et homotopie
Vendredi 21 juin 2019, 14 heures, Salle 1007
Alain Prouté Qu'est-ce qu'un ensemble ?

Si, il y a 10.000 ans, il y avait un mot pour désigner un troupeau de moutons, ce mot voulait aussi dire «ensemble», même si cette notion d'ensemble était moins abstraite que celle que Bolzano, Dedekind et Cantor introduisirent à la fin du XIXe siècle. On pourrait croire qu'après la théorie des ensembles de Zermelo et Fraenkel, la question de savoir ce qu'est un ensemble était close. Ce n'est pas le cas comme le remarque par exemple Quine, auteur d'une théorie alternative, qui dit dans les années 60 que la question est toujours ouverte. Une nouvelle impulsion sera donnée par Lawvere avec sa théorie axiomatique de la catégorie des ensembles, et surtout par Lawvere et Tierney avec la notion de topos élémentaire. Cette dernière notion inspirera Volger, puis Lambek, qui définira la notion de «dogme». J'ai rebaptisé leur théorie en «théorie des ensembles de Volger-Lambek». J'en donnerai une description et j'expliquerai pourquoi elle est, à condition qu'elle soit un peu généralisée, la bonne théorie de formalisation non seulement des ensembles, mais de toute la mathématique telle que nous la pratiquons aujourd'hui.

Combinatoire énumérative et analytique
Lundi 24 juin 2019, 11 heures, Salle 3052
Carola Doerr (CNRS, LIP6 Sorbonne University) Evolutionary Algorithms – From Theory to Practice and Back

Most real-world optimization problems do not have an explicit problem formulation, but only allow to compute the quality of selected solution candidates. Solving such black-box optimization problems requires iterative procedures which use the feedback gained from previous evaluations to determine the strategy by which the next solution candidates are generated. Many black-box optimization algorithms, such as Simulated Annealing, Evolutionary Algorithms, Swarm Intelligence Algorithms, are randomized – making it very difficult to analyze their performances mathematically.

In the last 15 years, the theory of randomized black-box optimization has advanced considerably, and has contributed to efficient optimization by providing insights into the working principles of black-box optimization which are hard or impossible to obtain by empirical means. On the other hand, empirically-guided benchmarking has opened up new research directions for theoretical investigations.

In this presentation we will discuss the state of the art in the theory of randomized black-box optimization algorithms. As part of this critical survey we will also mention a number of open questions and connections to other fields of Computer Science.

Algorithmes et complexité
Lundi 24 juin 2019, 11 heures, Salle 3052
Carola Doerr (CNRS, LIP6 Sorbonne University) Evolutionary Algorithms – From Theory to Practice and Back

Most real-world optimization problems do not have an explicit problem formulation, but only allow to compute the quality of selected solution candidates. Solving such black-box optimization problems requires iterative procedures which use the feedback gained from previous evaluations to determine the strategy by which the next solution candidates are generated. Many black-box optimization algorithms, such as Simulated Annealing, Evolutionary Algorithms, Swarm Intelligence Algorithms, are randomized – making it very difficult to analyze their performances mathematically.

In the last 15 years, the theory of randomized black-box optimization has advanced considerably, and has contributed to efficient optimization by providing insights into the working principles of black-box optimization which are hard or impossible to obtain by empirical means. On the other hand, empirically-guided benchmarking has opened up new research directions for theoretical investigations.

In this presentation we will discuss the state of the art in the theory of randomized black-box optimization algorithms. As part of this critical survey we will also mention a number of open questions and connections to other fields of Computer Science.

Systèmes complexes
Lundi 24 juin 2019, 11 heures, Salle 3052
Carola Doerr (CNRS, LIP6 Sorbonne University) Evolutionary Algorithms – From Theory to Practice and Back

Most real-world optimization problems do not have an explicit problem formulation, but only allow to compute the quality of selected solution candidates. Solving such black-box optimization problems requires iterative procedures which use the feedback gained from previous evaluations to determine the strategy by which the next solution candidates are generated. Many black-box optimization algorithms, such as Simulated Annealing, Evolutionary Algorithms, Swarm Intelligence Algorithms, are randomized – making it very difficult to analyze their performances mathematically.

In the last 15 years, the theory of randomized black-box optimization has advanced considerably, and has contributed to efficient optimization by providing insights into the working principles of black-box optimization which are hard or impossible to obtain by empirical means. On the other hand, empirically-guided benchmarking has opened up new research directions for theoretical investigations.

In this presentation we will discuss the state of the art in the theory of randomized black-box optimization algorithms. As part of this critical survey we will also mention a number of open questions and connections to other fields of Computer Science.

Graphes
Lundi 24 juin 2019, 11 heures, Salle 3052
Carola Doerr (CNRS, LIP6 Sorbonne University) Evolutionary Algorithms – From Theory to Practice and Back

Most real-world optimization problems do not have an explicit problem formulation, but only allow to compute the quality of selected solution candidates. Solving such black-box optimization problems requires iterative procedures which use the feedback gained from previous evaluations to determine the strategy by which the next solution candidates are generated. Many black-box optimization algorithms, such as Simulated Annealing, Evolutionary Algorithms, Swarm Intelligence Algorithms, are randomized – making it very difficult to analyze their performances mathematically.

In the last 15 years, the theory of randomized black-box optimization has advanced considerably, and has contributed to efficient optimization by providing insights into the working principles of black-box optimization which are hard or impossible to obtain by empirical means. On the other hand, empirically-guided benchmarking has opened up new research directions for theoretical investigations.

In this presentation we will discuss the state of the art in the theory of randomized black-box optimization algorithms. As part of this critical survey we will also mention a number of open questions and connections to other fields of Computer Science.