Mars 2017


Retour à la vue des calendrier
Jeudi 2 Mars
Heure: 12:30 - 13:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: On big data, optimization and learning
Description: Prof. Andrea Lodi In this talk I review a couple of applications on Big Data that I personally like and I try to explain my point of view as a Mathematical Optimizer -- especially concerned with discrete (integer) decisions -- on the subject. I advocate a tight integration of Machine Learning and Mathematical Optimization (among others) to deal with the challenges of decision-making in Data Science. For such an integration I try to answer three questions: 1) what can optimization do for machine learning? 2) what can machine learning do for optimization? 3) which new applications can be solved by the combination of machine learning and optimization?
Heure: 14:00 - 15:00
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Reachability Analysis of Pushdown Systems with an Upper Stack
Description: Adrien Pommellet Pushdown systems (PDSs) are a natural model for sequential programs, but they can fail to accurately represent the way an assembly stack actually operates. Indeed, one may want to access the part of the memory that is below the current stack or base pointer, hence the need for a model that keeps track of this part of the memory. To this end, we introduce pushdown systems with an upper stack (UPDSs), an extension of PDSs where symbols popped from the stack are not destroyed but instead remain just above its top, and may be overwritten by later push rules.

We prove that the sets of successors post* and predecessors pre* of a regular set of configurations of such a system are not always regular, but that post* is context-sensitive, so that we can decide whether a single configuration is forward reachable or not. In order to underapproximate pre* in a regular fashion, we consider a bounded-phase analysis of UPDSs, where a phase is a part of a run during which either push or pop rules are forbidden. We then present a method to overapproximate post* that relies on regular abstractions of runs of UPDSs. Finally, we show how these approximations can be used to detect stack overflows and stack pointer manipulations with malicious intent.
Vendredi 3 Mars
Heure: 11:00 - 12:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Introduction à la théorie de la complexité géométrique, d'après K. Mulmuley
Description: Luc Pellissier La théorie de la complexité géométrique est un programme de recherche porté par
Ketan Mulmuley, qui vise à résoudre des questions de complexité après les avoir
traduites comme des inclusions de surfaces algébriques représentant des groupes
de symétries.

Après avoir présenté la théorie avec beaucoup de recul, on présentera un
résultat de séparation obtenu ainsi (entre P et une classe ad hoc).
Lundi 6 Mars
Heure: 14:00 - 15:00
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Automatic Deception Detection in Text Applying Topic Modeling Algorithms
Description: Hiram Calvo We deal with deceptive text identification by using different kinds of features: a continuous semantic space model based on latent Dirichlet allocation topics (LDA), one-hot representation (OHR), syntactic information from syntactic n-grams (SN), and lexicon-based features using the linguistic inquiry and word count dictionary (LIWC). We will present experiments with several combinations of these features were tested to assess the best source(s) for deceptive text identification aiming to present a state of the art performance. We conducted our tests on three different available corpora: a corpus consisting of 800 reviews about hotels, a corpus consisting of 600 reviews about controversial topics, and a corpus consisting of 236 book reviews. Additionally, we present an analysis on which features lead to either deceptive or truthful texts, finding that certain words can play different roles (sometimes even opposing ones) depending on the task being evaluated. We will present results of experiments in one-domain setting by training and testing our models separately on each dataset (with fivefold cross-validation); in a mixed-domain setting by merging all datasets into one large corpus (again, with fivefold cross-validation), and finally, with cross-domain setting: using one dataset for testing and a concatenation of all other datasets for training.
Mardi 7 Mars
Heure: 11:30 - 12:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Valid quadratic inequalities for convex and some non-convex quadratic sets
Description: Julio César Góez In recent years, the generalization of Balas disjunctive cuts for mixed integer linear optimization problems to mixed integer non-linear optimization problems has received significant attention. Among these studies, mixed integer second order cone optimization (MISOCO) is a special case. For MISOCO one has the disjuncti ve conic cuts approach. That generalization introduced the concept of disjunctive conic cuts (DCCs) and disjunctive cylindrical cuts (DCyCs). Specifically, it showed that under some mild assumptions the intersection of those DCCs and DCyCs with a closed convex set, given as the intersection of a second order cone and an affine set, is the convex hull of the intersection of the same set with a linear disjunction. The key element in that analysis is the use of pencils of quadrics to find close forms for deriving the DCCs and DCyCs. In this talk we present an overview of the DCCs main results and we use the same approach to show the existence of valid conic inequalities for hyperboloids and non-convex quadratic cones when the disjunction is defined by parallel hyperplanes. Joint work with Miguel F. Anjos.
Jeudi 9 Mars
Heure: 10:30 - 11:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Hybride Modelling, Analysis and Quantitative Verification of Large Biological Regulatory Networks
Description: Louis Fippo Fitime Biological Regulatory Networks (BRNs) are usually used in systems biology for modelling,
understanding and controlling the dynamics of different biological functions (differentiation,
proliferation, proteins synthesis, apoptose) inside cells. Those networks are enhanced with
experimental data that are nowadays more available which give an idea on the dynamics of BRNs components.
Formal analysis of such models fails in front of the combinatorial explosion of generated behaviours
despite the fact that BRNs provide abstract representation of biological systems.

This thesis handles hybrid modelling, the simulation, the formal verification and control of Large
Biological Regulatory Networks. This modelling is done thanks to stochastic automata networks, thereafter
to Process Hitting by integrating time-series data.

Firstly, this thesis proposes a refining of the dynamics by estimation of stochastic and temporal (delay)
parameters from time-series data and integration of those parameters in automata networks models. This
integration allows the parametrisation of the transitions between the states of the system. Then,
a statistical analysis of the traces of the stochastic simulation is proposed to compare the dynamics
of simulations with the experimental data.

Secondly, this thesis develops static analysis by abstract interpretation in the automata networks
allowing efficient under- and over-approximation of quantitative (probability and delay) reachability properties.
This analysis enables to highlight the critical components to satisfy these properties.

Finally, taking advantage from the previous developed static analyses for the reachability properties in the
qualitative point of view, and from the power of logic programming (Answer Set Programming), this thesis addresses the domain
of control of system by proposing the identification of bifurcation transitions. Bifurcations are
transitions after which the system can no longer reach a state that was previously reachable.
Vendredi 10 Mars
Heure: 11:00 - 12:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Retrofitting linear types
Description: Arnaud Spiwack Type systems based on linear logic and their friends have proved that they are both capable of expressing a wealth of interesting abstractions. Among these the ability to mix garbage-collected and explicitly managed data in the same language. This is of prime interest for distributed computations that need to reduce latency induced by GC pauses: a smaller GC heap is a happier GC heap.

I had always had the belief that to add linear types to a language was a rather intrusive procedure and that a language with linear types would basically be an entirely new language. The Rust language is a case in point: it has a linear-like type system, but it's a very different language from your run-of-the-mill functional language.

This turns out not to be the case: this talk presents a way to extend a functional programming language. We are targeting Haskell but there is little specific to Haskell in this presentation. I will focus mostly on the type system and how it can be generalised: among other things the type system extends to dependent types.
Mardi 14 Mars
Heure: 14:00 - 17:00
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Géométrie combinatoire : densité d'hypergraphes et méthode probabiliste
Description: Xavier Goaoc Un hypergraphe à n sommets dont aucune projection sur k sommet n'est complète a au plus O(n^{k-1}) arêtes. Ce résultat, découvert simultanément par Sauer, Vapnick-Chervonenkis et Perles-Shelah au début des années 1970, est fondamental en apprentissage. En géométrie combinatoire et algorithmique, il permet souvent de contrôler la complexité (globale) d'une structure par sa "densité" locale. J'introduirai à ce mécanisme avant de présenter quelques extensions récentes, obtenues conjointement avec Boris Bukh (https://arxiv.org/abs/1701.06632), qui permettent un controle plus fin de la complexité. L'exposé ne supposera aucune connaissance préalable et un des ingrédients sera une construction probabiliste.
Jeudi 16 Mars
Heure: 15:30 - 16:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Preserving Partial Order Runs in Parametric Time Petri Nets
Description: Cesar Rodriguez Parameter synthesis for timed systems aims at deriving parameter
valuations satisfying a given property.
In this paper we target concurrent systems; it is well known that
concurrency is a source of state-space explosion, and partial order
techniques were defined to cope with this problem.
Here we use partial order semantics for parametric time Petri nets as a
way to significantly enhance the result of an existing synthesis
algorithm.
Given a reference parameter valuation, our approach synthesizes other
valuations preserving, up to interleaving,
the behavior of the reference parameter valuation.
We show the applicability of our approach using acyclic asynchronous circuits.

Join work with Thomas Chatain and Étienne André.
Vendredi 17 Mars
Heure: 11:00 - 12:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Des preuves, oui, mais formelles !
Description: Micaela MAYERO Dans cet exposé, très informel, lui, je parlerai de l'évolution des outils de preuves formelles ces 20, voire ces 30, dernières années. Nous nous appuierons sur quelques exemples afin de montrer comment les formalisations contribuent à la fois au développement de ces outils via des avancés théoriques majeurs ainsi qu'aux domaines qui s'engagent peu à peu dans leur utilisation. Suite à une première partie accessible à tout public, nous parlerons de la logique sous-jacente à l'un des système connaissant le plus d'avancées et de changements: Coq. Nous aborderons les évolutions de la théorie des types avec (im)prédicativité, la logique classique avec le choix de la totalité et d'autres avancées, pour finir, si nous avons le temps, par quelques mots sur CoqHoTT et l'axiome d'univalence.
Jeudi 23 Mars
Heure: 14:00 - 15:00
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: CARET Model Checking For Pushdown Systems
Description: Huu-Vu Nguyen CARET (A temporal logic of calls and returns) was introduced by Alur et al. This logic allows to write linear temporal logic formulas while taking into account matching of calls and returns. However, CARET model checking for Pushdown Systems (PDSs) was never considered in the literature. Previous works only dealt with the model checking problem for Recursive State Machine (RSMs). While RSMs are a good formalism to model sequential programs written in structured programming languages like C or Java, they become non suitable for modeling binary or assembly programs, since, in these programs, explicit push and pop of the stack can occur. Thus, it is very important to have a CARET model checking algorithm for PDSs. We tackle this problem in this paper. We also consider CARET model checking with regular valuations, where the set of configurations in which an atomic proposition holds is a regular language. We reduce these problems to the emptiness problem of Büchi Pushdown Systems. We implemented our technique in a tool, and we applied it to different case studies. Our results are encouraging. In particular, we were able to apply our tool to detect several malwares.

This is a joint work with Tayssir Touili.
Vendredi 24 Mars
Heure: 11:00 - 12:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Syntaxe transcendentale: seconde lecture.
Description: Christophe Fouqueré Suite de la première lecture: JY Girard, avec ses 3 articles sur la "syntaxe transcendentale", aborde la logique sous un angle à la fois philosophico-logique et sous un angle technique, pour aller au-delà de ce qui est fait jusqu'à présent: non seulement la logique linéaire propositionnelle mais encore le traitement de l'égalité donc du premier ordre. Modestement, je commencerai par présenter la partie technique, qui reprend et étend le modèle des réseaux de preuves, en présentant la représentation de MALL et des exponentielles.
Mardi 28 Mars
Heure: 14:00 - 17:00
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Phase Transition Threshold for Random Graphs and 2-SAT using Degree Constraints
Description: Sergey Dovgal We show that by restricting the degrees of the vertices of a graph to an arbitrary set ?, the threshold ?(?) of the phase transition for a random graph with n vertices and m = ?(?).n edges can be either accelerated (e.g.,?(?) approx 0.38 for ? = {0,1,4,5}) or postponed (e.g., ?(?) approx 0.95 for ?={1,2,50}) compared to a classical Erd?s?Rényi random graph where ?(N)=1/2. We investigate different graph statistics inside the critical window of transition (planarity, diameter, longest path...).We apply our results to a 2-SAT model with restricted literal degrees: the number of clauses that each literalis incident to belongs to the set ?. We prove a lower bound for the probability that a formula with n variables and m=2.?(?) n clauses is satisfiable. This probability is close to 1 for the subcritical regime m=2.?.n.(1-μ.n-1/3), μ to ∞ and improves/generalizes the lower bound of Bollobás, Borgs, Chayes, Kim, and Wilson.This shows how the phase transition threshold for 2-SAT moves if we change the degrees of the literals.Joint work with Vlady Ravelomanana.
Vendredi 31 Mars
Heure: 11:00 - 12:30
Lieu: Salle B107, bâtiment B, Université de Villetaneuse
Résumé: Équivalences entre programmes
Description: Jean-Yves Moyen Partant du théorème de Rice, on définit l'équivalence extensionelle comme "deux programmes sont équivalents si ils calculent la même fonction." Le théorème de Rice stipule alors une indécidabilité forte de cette équivalence (aucune union de classes n'est décidable).

Qu'en est-il des autres équivalences entre programmes ? Certaines sont décidables (eg, avoir le même nombre de lignes de codes), d'autres non. L'ensemble des équivalences possède une structure de treillis complet et le théorème de Rice parle de l'indécidabilité d'un filtre principal de ce treillis.

À partir de ces constatations, j'explore deux pistes parallèles. D'une part, quelle est l'importance de l'équivalence extensionelle dans ces résultats ? Est-ce qu'on peut obtenir d'autres résultats similaires à partir d'une autre équivalence ? D'autre part, comment étudier l'ensemble du treillis des équivalences ? Est-ce qu'on peut trouver des sous-ensembles qui conservent la structure lié à l'ordre tout en étant plus facilement manipulables (notamment, dénombrables) ?