Annonce March 2023

annonce@lrde.epita.fr

4 participants
6 discussions

Accepted paper at TUG 2023

by Didier Verna

Dear all, I'm happy to announce that the following paper has been accepted at the TUG 2023 conference, Bonn, July 14-16 2023. Interactive and Real-Time Typesetting for Demonstration and Experimentation Abstract: In general, typesetting experimentation is not a very practical thing to do. WYSIWYG typesetting systems are very reactive but do not offer highly configurable algorithms, and TeX, with its separate development / compilation / visualization phases, is not as interactive as its WYSIWYG competitors. Being able to experiment with typesetting algorithms interactively and in real-time is nevertheless desirable, for instance for demonstration purposes, or for rapid prototyping and debugging of new ideas. We present ETAP (Experimental Typesetting Algorithms Platform), a tool written to ease typesetting experimentation and demonstration. ETAP currently provides several paragraph justification algorithms, all with many configuration options such as kerning, ligatures, flexible spaces, sloppiness, hyphenation, etc. The resulting paragraph is displayed with many visual hints as well, such as paragraph, character, and line boxes, baselines, over/underfullness hints, hyphenation clues, etc. All these parameters, along with the desired paragraph width, are adjustable interactively through a GUI, and the resulting paragraph is displayed and updated in real-time. But ETAP can also be used without, or in conjunction with the GUI, as a scriptable application. In particular, it is able to generate all sorts of statistical reports or charts on the behavior of the various algorithms, for instance, the number of over/underfull boxes per paragraph width, the average compression or stretch ratio per line, whatever else you want. This allows you to quickly demonstrate or evaluate the comparative behavior or merits of the provided algorithms, or whichever you may want to add to the pool. -- Resistance is futile. You will be jazzimilated. Lisp, Jazz, Aïkido: http://www.didierverna.info

2 years

Séminaire IMAGE mer. 15/03 : Parsing 2D pour l'analyse de documents (B. Coüasnon)

by Joseph Chazalon

Bonjour à tous, je suis heureux de vous inviter au séminaire du mercredi 15 mars prochain, qui aura lieu de 11h à 12h en amphi 0 (Paris, KB). Bertrand était mon encadrant de doctorat et j'ai travaillé sur le système qu'il présentera (qui a bien évolué depuis). Je pense qu'il pourra intéresser les curieux de l'équipe automates aussi, pour l'usage assez extrême des compilateurs et des grammaires qui est fait (grammaire bi-dimensionnelles avec un analyseur codé en lambda prolog qui a la capacité de déplacer non linéairement sa tête de lecture sur la surface à analyser en fonction du contexte passé). Si la récursion gauche ne vous fait pas peur, venez jeter un œil. Je pense qu'il pourrait aussi intéresser les curieux côté IA car ma motivation derrière cette invitation est de démarrer un travail visant à bricoler l'architecture des décodeurs de transformeurs pour améliorer les deep parsers dont les premiers résultats sont assez époustouflants. Et bien entendu, pour ceux qui s'intéressent à l'analyse d'images (de documents), ce système est un cas d'école. Lien Teams pour celles et ceux à distance : https://teams.microsoft.com/l/meetup-join/19%3apzIZ9vVKzWORw57NZInOVyAH2uKq… En espérant vous voir nombreux, Joseph ======================================================================= Invité : Bertrand Coüasnon (INSA Rennes/IRISA) Titre : DMOS-PI : un système complet d’analyse interactive de collections de documents Résumé : DMOS-PI est un système de reconnaissance de collections de documents qui s’appuie sur une description visuelle des documents grâce à des grammaires bidimensionnelles, combinées à des méthodes d’apprentissage profond. Il permet également la modélisation de collections de documents, une formalisation d’interactions asynchrones entre l’analyse des pages et les saisies manuelles, et une transformation des informations unitaires rencontrées dans les pages en des séquences de ces informations unitaires, le tout étant intégré progressivement par une analyse itérative des pages. Cette combinaison produit une fiabilisation des informations reconnues s’appuyant sur les redondances rencontrées dans les collections de documents et permet d’envisager la construction de systèmes auto-adaptatifs. Ce système a été validé sur une très grande variété de documents : des formulaires administratifs anciens, de la presse ancienne, des documents récents dégradés, des listes de cotations boursières du XIXe et XXe siècle, des partitions musicales anciennes… =======================================================================

2 years

Soutenance de thèse Yizi CHEN 22 Mars 2023 à l'IGN

by Yizi Chen

Dear LRE lab members: I am pleased to invite you to attend my Ph.D. defense on 22 March. The defense will start at 2:00 PM and be held at IGN (room BAT A 1er étage pièce 182 François ARAGO, 73 Av. de Paris, 94160 Saint-Mandé). For security reasons, please register using the following form if you plan to attend my defense: [ https://docs.google.com/forms/d/e/1FAIpQLSdDNayv13-XdRFy1LP3uX-yeHK4OJekByU… | https://docs.google.com/forms/d/1IDVQmlM4lLPFLH7EwHZxn_fxygvqCxRCPKMgJhHYTQ… ] The title and abstract of the Ph.D. dissertation are available below: Thesis title: Modern vectorization and alignment of historical maps: An application to Paris Atlas (1789-1950) Abstract: Maps have been a unique source of knowledge for centuries. Such historical documents provide invaluable information for analyzing complex spatial transformations over important time frames. This is particularly true for urban areas that encompass multiple interleaved research domains: humanities, social sciences, etc. The large amount and significant diversity of map sources call for automatic image processing techniques in order to extract the relevant objects as vector features. The complexity of maps (text, noise, digitization artifacts, etc.) has hindered the capacity of proposing versatile and efficient raster-to-vector approaches for decades. In this thesis, we propose a learnable, reproducible, and reusable solution for the automatic transformation of raster maps into vector objects (building blocks, streets, rivers), focusing on the extraction of closed shapes. Our approach is built upon the complementary strengths of convolutional neural networks which excel at filtering edges while presenting poor topological properties for their outputs, and mathematical morphology, which offers solid guarantees regarding closed shape extraction while being very sensitive to noise. In order to improve the robustness of deep edge filters to noise, we review several and propose new topology-preserving loss functions which enable to improve of the topological properties of the results. We also introduce a new contrast convolution (CConv) layer to investigate how architectural changes can impact such properties. Finally, we investigate the different approaches which can be used to implement each stage, and how to combine them in the most efficient way. Thanks to a shape extraction pipeline, we propose a new alignment procedure for historical map images, and start to leverage the redundancies contained in map sheets with similar contents to propagate annotations, improve vectorization quality, and eventually detect evolution patterns for later analysis or to automatically assess vectorization quality. To evaluate the performance of all methods mentioned above, we released a new dataset of annotated historical map images. It is the first public and open dataset targeting the task of historical map vectorization. We hope that thanks to our publications, public and open releases of datasets, codes, and results, our work will benefit a wide range of historical map-related applications. Thanks in advance! Yizi Chen

2 years

[CfPart] 16th European Lisp Symposium, April 24-25, Amsterdam

by Didier Verna

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 16th European Lisp Symposium In-Cooperation-With: ACM SIGLAN Call for Participation April 24-25, 2023 Startup Village, Amsterdam, Nederlands https://www.european-lisp-symposium.org/2023 Sponsored by EPITA, DIRO, MLPrograms, Franz Inc., and SISCOG ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ Recent News ~~~~~~~~~~~ Registrations are now open (early bird deadline: April 9) Keynote speakers announced (see below) Important Dates ~~~~~~~~~~~~~~~ - Author notification: March 26, 2023 - Final papers due: April 9, 2023 - Early registration deadline: April 9, 2023 - Symposium: April 24-25, 2023 Scope ~~~~~ The European Lisp Symposium is a premier forum for the discussion and dissemination of all aspects of design, implementation, and application of any of the Lisp dialects, including Common Lisp, Scheme, Emacs Lisp, Clojure, Racket, ACL2, AutoLisp, ISLISP, Dylan, SKILL, Hy, Shen, Carp, Janet, uLisp, Picolisp, Gamelisp, TXR, and so on. We encourage everyone interested in Lisp to participate. The European Lisp Symposium invites high quality papers about novel research results, insights and lessons learned from practical applications, and educational perspectives. We also encourage submissions about known ideas as long as they are presented in a new setting and/or in a highly elegant way. Topics include but are not limited to: - context-, aspect-, domain-oriented and generative programming - macro-, reflective-, meta- and/or rule-based development approaches - language design and implementation - language integration, inter-operation, and deployment - development methodologies, support, and environments - educational approaches and perspectives - experience reports and case studies Keynotes ~~~~~~~~ ##### Artificial Intelligence: a Problem of Plumbing? -- Gerald J. Sussman, MIT CSAIL, USA We have made amazing progress in the construction and deployment of systems that do work originally thought to require human-like intelligence. On the symbolic side we have world-champion Chess-playing and Go-playing systems. We have deductive systems and algebraic manipulation systems that exceed the capabilities of human mathematicians. We are now observing the rise of connectionist mechanisms that appear to see and hear pretty well, and chatbots that appear to have some impressive linguistic ability. But there is a serious problem. The mechanisms that can distinguish pictures of cats from pictures of dogs have no idea what a cat or a dog is. The chatbots have no idea what they are talking about. The algebraic systems do not understand anything about the real physical world. And no deontic logic system has any idea about feelings and morality. So what is the problem? We generally do not know how to combine systems so that a system that knows how to solve problems of class A and another system that knows how to solve problems of class B can be combined to solve not just problems of class A or class B but can solve problems that require both skills that are needed for problems of class A and skills that are needed for problems of class B. Perhaps this is partly a problem of plumbing. We do not have linguistic structures that facilitate discovering and building combinations. This is a fundamental challenge for the programming-language community. We need appropriate ideas for abstract plumbing fittings that enable this kind of cooperation among disparate mechanisms. For example, why is the amazingly powerful tree exploration mechanism that is used for games not also available, in the same system, to a deductive engine that is being applied to a social interaction problem? I will attempt to elucidate this problem and perhaps point at avenues of attack that we may work on together. ##### Gradual, Multi-Lingual, and Teacher-Centric Programming Education -- Felienne Hermans, Vrije Universiteit Amsterdam, Nederlands (tba) Programme Chair ~~~~~~~~~~~~~~~ Stefan Monnier, DIRO, Université de Montréal, Canada Programme Committee ~~~~~~~~~~~~~~~~~~~ Mark Evenson, not.org, Austria Marco Heisig, Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany Ioanna Dimitriou, Igalia S.L., Germany Robert Smith, HRL Laboratories Mattias Engdegård Marc Feeley, Université de Montréal, Canada Marc Battyani, FractalConcept Alan Ruttenberg, National Center for Ontological Research, USA Nick Levine, Ravenbrook Ltd, UK Ludovic Courtès, Inria, France Matthew Flatt, University of Utah, USA Irène Durand, Université Bordeaux 1, France Jay McCarthy, Brigham Young University, USA Ambrose Bonnaire-Sergeant, Cisco Christopher League, Long Island University, NY, USA Pascal Costanza, Intel, Belgium Christian Queinnec Local Chair ~~~~~~~~~~~ Breanndán Ó Nualláin, Machine Learning Programs, Nederlands -- Resistance is futile. You will be jazzimilated. Lisp, Jazz, Aïkido: http://www.didierverna.info

2 years, 1 month

Paper published in Information & Computation

by Uli Fahrenberg

The following paper has been published in Information & Computation: Posets With Interfaces as a Model for Concurrency Uli Fahrenberg, Christian Johansen, Georg Struth, Krzysztof Ziemiański We introduce posets with interfaces (iposets) and generalise their standard serial composition to a new gluing composition. In the partial order semantics of concurrency, interfaces and gluing allow modelling events that extend in time and across components. Alternativelytaking a decompositional view, interfaces allow cutting through events, while serial composition may only cut through edges of a poset. We show that iposets under gluing composition form a category, which generalises the monoid of posets under serial composition up to isomorphism. They form a 2-category when a subsumption order and a lax tensor in the form of a non-commutative parallel composition are added, which generalises the interchange monoids used for modelling series-parallel posets. We also study the gluing-parallel hierarchy of iposets, which generalises the standard series-parallel one. The class of gluing-parallel iposets contains that of series-parallel posets and the class of interval orders, which are well studied in concurrency theory, too. We also show that it is strictly contained in the class of all iposets by identifying several forbidden substructures. https://www.lrde.epita.fr/wiki/Publications/fahrenberg.22.iandc

2 years, 1 month

Paper accepted at FM 2023: Energy Problems in Finite and Timed Automata with Büchi Conditions

by Uli Fahrenberg

We are happy to announce that the following article has been accepted at the he 25th International Symposium on Formal Methods (FM 2023) to be held in Lübeck in March 2023: Energy Problems in Finite and Timed Automata with Büchi Conditions Sven Dziadek, Uli Fahrenberg and Philipp Schlehuber-Caissier Abstract: We show how to efficiently solve energy Büchi problems in finite weighted Büchi automata and in one-clock weighted timed Büchi automata; all our algorithms are implemented in a pipeline based on TChecker and Spot. Solving the latter problem is done by using the corner-point abstraction; the former problem is handled by a modified version of Bellman-Ford interleaved with Couvreur's algorithm. The paper is available at https://arxiv.org/abs/2205.04392

2 years, 1 month

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Annonce March 2023