BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//LORIA - ECPv6.17.4.1//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:LORIA
X-ORIGINAL-URL:https://www.loria.fr
X-WR-CALDESC:Évènements pour LORIA
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:Europe/Paris
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20170326T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20171029T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20180325T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20181028T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20190331T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20191027T010000
END:STANDARD
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:20200329T010000
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:20201025T010000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190423T110000
DTEND;TZID=Europe/Paris:20190423T120000
DTSTAMP:20190417T152232Z
CREATED:20190417T152225Z
LAST-MODIFIED:20190417T152232Z
UID:7022-1556017200-1556020800@www.loria.fr
SUMMARY:MALOTEC Seminar
DESCRIPTION:For the next MALOTEC seminar\, Valia Mitsou\, who is a candidate for the position of associate professor in CentraleSupelec will give a presentation at LORIA on Tuesday April 23 at 11am in room A008 \nPresentation of the speaker : \nValia Mitsou\nUniversité Paris Diderot\nInstitut de Recherche en Informatique Fondamentale (IRIF) \nTitle: Limitation of treewidth for problems beyond NP \nAbstract: In this seminar\, we take a closer look at the parameterized complexity of problems belonging in Σ^p_2 and Π^p_2\, the second level of the polynomial hierarchy. We provide tight fine-grained bounds on their complexity with respect to the most important structural graph parameter\, the treewidth. We observe that these problems exhibit similar behavior: we show that a variety of diverse problems including ∃∀SAT\, Choosability\, as well as various problems from AI such as Abduction and Abstract Argumentation\, while they admit a $2^{2^{O(tw)}}$ algorithm\, they cannot be solved in time $2^{2^{o(tw)}}$ under the Exponential Time Hypothesis. \nFor more information
URL:https://www.loria.fr/event/malotec-seminar-5/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190404T140000
DTEND;TZID=Europe/Paris:20190404T150000
DTSTAMP:20190402T133202Z
CREATED:20190402T133158Z
LAST-MODIFIED:20190402T133202Z
UID:6900-1554386400-1554390000@www.loria.fr
SUMMARY:MALOTEC seminar : Lydia Boudjeloud-Assala
DESCRIPTION:  \n=========================================================\nMALOTEC Seminar: Thursday April 04 2019\, 2pm\, Room C005 \nLydia Boudjeloud-Assala\nUniversité de Lorraine\nLoria \nTitle: Human in the loop: Visual -massive and temporal- data mining \nAbstract: I will present a review of my research on cooperative\napproaches combining interactive visualization methods and automatic\nmethods (attributes Selection\, Clustering\, Biclustering\, and outlier\ndetection) for the knowledge discovery in massive and temporal data. I\nwill present different possibilities to involve the user in the data\nmining process in order to improve his confidence and understanding in\nthe models or obtained results. New challenges for these methods\, which\nmust be able to handle not only increasing amounts of data but also data\nthat can change over time\, will also be presented and discussed. \nFor more information \nFurther MALOTEC Seminars \n  \n 
URL:https://www.loria.fr/event/malotec-seminar-lydia-boudjeloud-assala/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190404T133000
DTEND;TZID=Europe/Paris:20190404T150000
DTSTAMP:20190305T074902Z
CREATED:20190305T074902Z
LAST-MODIFIED:20190305T074902Z
UID:6596-1554384600-1554390000@www.loria.fr
SUMMARY:SSL Seminar
DESCRIPTION:Next SSL seminar will take place on Thursday\, April 4th at 1.30 pm in room A008. \nGilles Barthe (MPI Bochum) will give a presentation entitled « Computer-aided cryptography ». \n \nWe need cryptography that we can trust. Yet the design\, analysis\, and implementation of cryptographic libraries is a challenging task\, that requires insights across various areas of mathematics and computer science. Computer-aided cryptography is a young research area which\nuses formal methods for exploring the design space of cryptographic constructions and for delivering zero-defect\, side-channel resistant\, cryptographic libraries. The talk will give an overview of our work in computer-aided cryptography and discuss some of the challenges and opportunities in cryptography and beyond.
URL:https://www.loria.fr/event/ssl-seminar-8/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190402T100000
DTEND;TZID=Europe/Paris:20190402T110000
DTSTAMP:20190329T165913Z
CREATED:20190329T142924Z
LAST-MODIFIED:20190329T165913Z
UID:6809-1554199200-1554202800@www.loria.fr
SUMMARY:Séminaire D3: Ingénierie multi-agent des collectifs cyber-physiques
DESCRIPTION:Le prochain séminaire D3 aura lieu le mardi 2 avril à 10h en salle A008. \nVoici le titre et le résumé: \nTitre : Ingénierie multi-agent des collectifs cyber-physiques (par Jean-Paul Jamont – Univ. Grenoble Alpes) \nRésumé : Les collectifs cyber-physiques sont des systèmes embarqués en réseau dans lequel les nœuds logiciels et matériels ont une autonomie de décision et coopèrent spontanément afin de participer à l’accomplissement d’objectifs du système global ou de pallier des manques de connaissances ou de compétences individuelles. Ces objectifs portent notamment sur l’état de leur environnement physique. Cette présentation introduit les nombreux défis posés par leur conception et propose une ingénierie de ces collectifs basée sur une démarche\, des modèles et des outils multi-agents. Plus loin l’ingénierie de leur couplage permettra d’aborder la notion de systèmes de systèmes cyber-physiques.
URL:https://www.loria.fr/event/seminaire-d3-ingenierie-multi-agent-des-collectifs-cyber-physiques/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190402T093000
DTEND;TZID=Europe/Paris:20190402T160000
DTSTAMP:20190329T153249Z
CREATED:20190329T153245Z
LAST-MODIFIED:20190329T153249Z
UID:6811-1554197400-1554220800@www.loria.fr
SUMMARY:Journée scientifique de l'équipe ABC
DESCRIPTION:Programme de la journée scientifique de l’équipe ABC\nRV Salle C 103 du bâtiment LORIA le 2 avril 2019 ! \nProgramme : \n9h30 – 10h00 –  Accueil des participants\n10h – 11h00  – Parisa Rastin\, Université Paris 13 Apprentissage automatique et adaptatif pour le clustering de flux de données relationnelles\n11h00 – 11h15 Pause\n11h15 – 12h15 Marianne Clausel\, Université de Lorraine Topic modeling\n12h15 – 12h35 Myriam Maumy-Bertrand\, Université de Strasbourg Sur quelques aspects du dilemme du Petit Chaperon rouge : discrimination entre la mère-grand et le grand méchant loup\n12h35 – 13h30 Repas\n13h30 – 14h30 Myriam Tami\, CentraleSupélec Uncertain Trees : Dealing with Uncertain Inputs in Regression Trees\n14h30 – 16h00 Discussion sur l’élaboration d’un projet scientifique commun\n16h00 –  Fin de la journée
URL:https://www.loria.fr/event/journee-scientifique-de-lequipe-abc/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190327T110000
DTEND;TZID=Europe/Paris:20190327T120000
DTSTAMP:20190325T090116Z
CREATED:20190325T090107Z
LAST-MODIFIED:20190325T090116Z
UID:6773-1553684400-1553688000@www.loria.fr
SUMMARY:MALOTEC Seminar : Saïd Jabbour
DESCRIPTION:Next MALOTEC seminar will take place on Wednesday March 27th at 11am in room B013. \nSaïd Jabbour (Université d’Artois\, Centre de recherche en informatique de lens (CRIL)) will give a presentation entitled « Data mining by constraints ». \nAbstract: In this presentation\, I will address the issue of data mining\nby constraints. I will present some contributions on declarative\napproaches for different data mining tasks including frequent itemset\nmining and its various condensed forms\, association rules\, sequential\npattern mining\, uncertain frequent itemset mining and gradual itemset\nmining. To highlight the cross-fertilization between symbolic AI and\ndata mining\, I will show how linear constraints can be used to summarize\nlarge graphs\, how data mining techniques can be exploited to compress\nBoolean formulas and finally how the concept of symmetry\, widely\nexplored in SAT/CP\, is extended to pattern mining. \nFor more information: http://www.cril.univ-artois.fr/~jabbour/
URL:https://www.loria.fr/event/malotec-seminar-4/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190319T140000
DTEND;TZID=Europe/Paris:20190319T153000
DTSTAMP:20190313T105411Z
CREATED:20190313T105411Z
LAST-MODIFIED:20190313T105411Z
UID:6659-1553004000-1553009400@www.loria.fr
SUMMARY:Seminar : Adrien Boiret
DESCRIPTION:Adrien Boiret will give a presentation on Tuesday\, 19th March at 2pm in room A008. \nHis talk is entitled « Extend the RPNI algorithm for Symbolic Learning to Transducers with Lookahead ». \n \nAbstract:\nWe consider the problem of passive symbolic learning of tree and word transducers.\nSymbolic learning means learning an exact target (automaton\, transducer…) using exact information on its (words\, input/output pairs…).\nThe passive learning problem deals with identifying a specific transducer in normal form from a finite set of behaviour examples (as opposed to the active learning that uses student-teacher interactions). \nSymbolic learning problems are solved in word languages using the RPNI algorithm\, that relies heavily on the Myhill-Nerode characterization of a minimal normal form on DFA.\nThe extensions of RPNI to word transformations and tree languages follow the same pattern: first\, a Myhill-Nerode theorem is identified\, then the normal form it induces can be learnt from examples.\n\nIn this talk I will present another extension of the RPNI model that necessitates a departure from this schema to learn rational functions on words\, performed by transducers with lookahead.\nThe notable difference is that the nature of the minimal normal form on these transducers is ill adapted to classic RPNI methods\, and thus a new normal form has to be found.
URL:https://www.loria.fr/event/seminar-adrien-boiret/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190319T140000
DTEND;TZID=Europe/Paris:20190319T150000
DTSTAMP:20190319T085158Z
CREATED:20190315T142954Z
LAST-MODIFIED:20190319T085158Z
UID:6679-1553004000-1553007600@www.loria.fr
SUMMARY:Séminaire : Joël Legrand
DESCRIPTION:Joël Legrand\, MCF à Lille de l’équipe Inria Magnet\, viendra nous présenter ses travaux le mardi 19 mars à 14h en salle A007 à CentraleSupélec (2 rue Edouard Belin\, Metz). \nSa présentation est intitulée « Syntax-based Word Sequence Modeling using Deep Learning ». \n \nAbstract: \nTraditionally\, natural language processing (NLP) relied on generative models with task specific and manually engineered features.\nRecently deep learning approaches provided state of-the-art results in various fields such as computer vision\, speech processing and natural language processing.\nThe central idea behind these approaches is to learn features and models simultaneously\, in an end-to-end manner\, and making as few assumptions as possible.\nIn NLP\, word embeddings\, mapping words in a dictionary on a continuous low-dimensional vector space\, have proven to be very efficient for a large variety of tasks while requiring almost no a-priori linguistic assumptions. \nIn this talk\, I will present some results on representations of segments of sentences using deep neural network models.\nIn particular\, I will show how syntactic structures\, such as consistuency and dependency syntactic trees\, can be leveraged for the\npurpose of solving NLP tasks that involve complex sentence-level relationships.\nI will first introduce the key concepts of deep learning for NLP. I will then focus on two empirical studies concerning the tasks of\nsyntactic parsing using recursive neural networks (RNN) and relationship extraction from text using tree-LSTM with dependency structures. \nSelected publications:\n[1]. Joint RNN-based greedy parsing and word composition.\nJoël Legrand and Ronan Collobert.\nProceeding of the 3rd International Conference on Learning\nRepresentations (ICLR 2015)\nhttp://arxiv.org/pdf/1412.7028 \n[2] Syntax-based Transfer Learning for the Task of Biomedical Relation Extraction\nJoël Legrand\, Yannick Toussaint\, Chedy Raïssi\, Adrien Coulet\nProceeding of the Ninth International Workshop on Health Text Mining and\nInformation Analysis (LOUHI 2018)\nhttps://hal.inria.fr/hal-01869071/document \nShort bio:\nJoël Legrand received his MSc degree in Computer Science from the\nUniversité de Lorraine and his Ph.D in Electrical Engineering from the\nÉcole Polytechnique Fédérale de Lausanne (EPFL\, Switzerland).\nHe recently joined the  MAGNET team (Cristal\, INRIA-CNRS-Université de\nLille) as an associate professor (maître de conférences).
URL:https://www.loria.fr/event/seminaire-joel-legrand/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;VALUE=DATE:20190311
DTEND;VALUE=DATE:20190315
DTSTAMP:20190208T101437Z
CREATED:20190208T101437Z
LAST-MODIFIED:20190208T101437Z
UID:6527-1552262400-1552607999@www.loria.fr
SUMMARY:MLA 2019
DESCRIPTION:The Third Workshop on Mathematical Logic and its Applications will take place at Loria between March 11th and March 14th. \nMore information : mla2019.loria.fr \nWorkshop Description\nThe workshop brings together researchers of mathematical logic and its related areas\, and would provide a forum for interplay between these areas. This workshop\, organised by Loria\, will be held as a part of the JSPS Core-to-Core Program “Mathematical Logic and its Applications”\, which is led by Japan Advanced Institute of Science and Technology (JAIST) in cooperation with several institutes both in Japan and abroad. Anyone\, not restricted to the members of the project\, is welcome to participate. \nScope\nThe scope includes (but is not limited to) the following topics and their applications: \n\nIntuitionistic logic\nType theory\nProof theory\nConstructive analysis/topology\nProgram extraction\nReverse mathematics\nComputable analysis/topology\nRecursion theory (Computability theory)\nNon-classical logic\nComputational complexity\n\nSubmission\nExtended abstracts (1-2 pages) of original work are welcome.\nThe deadline is January 28\, 2019.\nUpload your submission via EasyChair. \nImportant dates\nSubmission: January 28\, 2019\nNotification: February 1st\, 2019\nDeadline for registration: February 20\, 2019\nWorkshop: March 11-14\, 2019
URL:https://www.loria.fr/event/mla-2019/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190308T133000
DTEND;TZID=Europe/Paris:20190308T150000
DTSTAMP:20190115T104422Z
CREATED:20190115T104422Z
LAST-MODIFIED:20190115T104422Z
UID:6383-1552051800-1552057200@www.loria.fr
SUMMARY:D3 Seminar : David Gutierrez Estevez\, Samsung
DESCRIPTION:Next D3 Seminar will take place on Friday\, March 8th at 1.30 pm in room A008. \nDavid Gutierrez Estevez (Samsung) will give a presentation entitled « 5G Architecture and Elastic Network Slicing: A View from the 5G-MoNarch Project ». \nAbstract:  Vertical markets and industries are addressing a large diversity of heterogeneous services\, use cases\, and applications in 5G. It is currently common understanding that for networks to be able to satisfy those needs\, a flexible\, adaptable\, and programmable architecture based on network slicing is required. Moreover\, a softwarization and cloudification of the networks is already underway\, where network functions (NFs) are transformed from monolithic pieces of equipment to programs running over a shared pool of computational and communication resources. In this talk\, an overview will be given on the topic of 5G architecture with a special emphasis on network slicing. Furthermore\, the talk will encompass the view on the topic from the 5G-PPP research project 5G-MoNArch (5G MobileNetwork Architecture for diverse services\, use cases\, and applications in 5G and beyond)\, which has been working for the last 18 months on designing an architecture that will combine today’s consolidated concepts (such as virtualization\, slicing and orchestration of access and core functions) with several novel innovations targeted at optimizing the envisioned architecture. In particular\, the talk will look at the network resource elasticity innovation\, a key means to make an efficient use of the computational resources in 5G systems aided by artificial intelligence and data analytics. \nBio: Dr. David M. Gutierrez Estevez obtained his Engineering Degree in Telecommunications (Hons.) from the Universidad de Granada\, Spain\, and his M.S. and Ph.D. degrees from Georgia Institute of Technology in Atlanta\, USA. He developed his Ph.D. thesis at the Broadband Wireless Networking Laboratory under the supervision of Prof. Ian F. Akyildiz\, obtaining the Researcher of the Year Award in 2013. From Sept. 2014 to Sept. 2015\, David worked for Huawei Technologies as Principal Research Engineer in Silicon Valley in the area of cloud computing. Previous to that\, David had held an internship position at the Corporate R&D Division of Qualcomm\, as well as research assistant and intern positions at Fraunhofer Heinrich Hertz Institute and Fraunhofer Institute for Integrated Circuits\, both in Germany. Since January 2016\, David is with the 5G Research and Standards team of the Samsung Electronics R&D Institute UK\, currently holding a Principal Engineer position where he leads several research projects and participates in different standardization activities related to network architecture. \n 
URL:https://www.loria.fr/event/d3-seminar-david-gutierrez-estevez-samsung/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190301T093000
DTEND;TZID=Europe/Paris:20190301T110000
DTSTAMP:20190218T131956Z
CREATED:20190218T131956Z
LAST-MODIFIED:20190218T131956Z
UID:6572-1551432600-1551438000@www.loria.fr
SUMMARY:Seminar : Xiaofang Wang
DESCRIPTION:Xiaofang Wang (Inria Rennes) will give a seminar on March 1st at 9.30 am in room A006. \nHer presentation is entitled « From Image segmentation and active colloids tracking\, to weakly supervised object detection ». \nI will present my research activities\, covering four axes in machine learning. They are unsupervised learning for image segmentation and active colloids tracking\, weakly supervsed object detection\, supervised learning for 3D object classification and domain adaptation in transfer learning. First\, a short overview on all axes is introduced. Then I will present several works in depth\, including graph based image segmentation and object tracking\, as well as what is weakly supervised in object detection and our contribution. Finally\, a theorectical introduction on the domain adaptation and our proposals will be discussed. Please find more details here:https://team.inria.fr/rainbow/fr/xiaofang-wang/.
URL:https://www.loria.fr/event/seminar-xiaofang-wang/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190228T133000
DTEND;TZID=Europe/Paris:20190228T150000
DTSTAMP:20190207T143456Z
CREATED:20190207T143456Z
LAST-MODIFIED:20190207T143456Z
UID:6519-1551360600-1551366000@www.loria.fr
SUMMARY:SSL Seminar : Aurore Guillevic
DESCRIPTION:Next SSL Seminar will take place on Thursday\, February 28th at 1.30 pm in room A008. \nAurore Guillevic (Caramba) will give a presentation entitled « Discrete logarithm computation in finite fields GF(p^k) with NFS variants and consequences in pairing-based cryptography. » \n \nAbstract:\nPairings on elliptic curves are involved in signatures\, NIZK\, and recently in blockchains (ZK-SNARKS).\nThese pairings take as input two points on an elliptic curve E over a finite field\, and output a value in an extension of that finite field.\nUsually for efficiency reasons\, this extension degree is a power of 2 and 3 (such as 12\,18\,24)\, and moreover the characteristic of the finite field has a special form. The security relies on the hardness of\ncomputing discrete logarithms in the group of points of the curve and in the finite field extension. \nIn 2013-2016\, new variants of the function field sieve and the number field sieve algorithms turned out to be faster in certain finite fields related to pairing-based cryptography. Now small characteristic settings\n(with GF(2^(4*n))\, GF(3^(6*m))) are discarded\, and the situation of GF(p^k) where p is prime and k is small (in practice from 2 to 54) is unclear.\nThe asymptotic complexity of the Number Field Sieve algorithm in finite fields GF(p^k) (where p is prime) and its Special and Tower variants is given by an asymptotic formula of the form A^(c+o(1)) where A depends on\nthe finite field size (log p^k)\, o(1) is unknown\, and c is a constant between 1.526 and 2.201 that depends on p\, k\, and the choice of parameters in the algorithm. \nIn this work we improve the approaches of Menezes-Sarkar-Singh and Barbulescu-Duquesne to estimate the cost of a hypothetical implementation of the Special-Tower-NFS in GF(p^k) for small k (k <= 24)\, and update some parameter sizes for pairing-based cryptography. \nThis is a joint work with Shashank Singh\, IISER Bhopal\, India. \nMore information about SSL Seminars
URL:https://www.loria.fr/event/ssl-seminar-aurore-guillevic/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190228T110000
DTEND;TZID=Europe/Paris:20190228T120000
DTSTAMP:20190227T110218Z
CREATED:20190227T110218Z
LAST-MODIFIED:20190227T110218Z
UID:6581-1551351600-1551355200@www.loria.fr
SUMMARY:MALOTEC Seminar : Marc Plantevit
DESCRIPTION:Next MALOTEC seminar will take place on Thursday\, February 28th at 11am in room A008. \nMarc Plantevit (Université Claude Bernard Lyon 1 – LIRIS – CNRS UMR 5205) will give a presentation entitled « Pattern Mining in Augmented Graphs ». \n  \nMarc Plantevit\nUniversité Claude Bernard Lyon 1\nData Mining and Machine Learning (DM2L)\nLIRIS – CNRS UMR 5205 \nTitle: Pattern Mining in Augmented Graphs \nAbstract: Graphs are a powerful mathematical abstraction that enables to depict many real world phenomena. Vertices describe entities and edges identify relations between entities. Such graphs are often augmented with additional pieces of information. For instance\, the vertices or the edges are enriched with attributes describing them and are called vertex (respectively edge) attributed graphs. Graphs can also be dynamic\, i.e.\, the structure and the values of vertex attributes may evolve through time. The discovery of patterns in such graphs may provide actionable insights and boost the user knowledge. In this talk\, I will discuss the different pattern domains for augmented graphs I contributed to define. This includes the discovery of exceptional attributed subgraphs in edge or vertex attributed graphs. Then\, I will discuss how to find patterns of higher interest by taking into account the domain knowledge\, user feedback and user’s prior knowledge through different examples. \nFor more information: https://perso.liris.cnrs.fr/mplantev/doku/
URL:https://www.loria.fr/event/malotec-seminar-marc-plantevit/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190129T110000
DTEND;TZID=Europe/Paris:20190129T233000
DTSTAMP:20190116T143344Z
CREATED:20190116T143344Z
LAST-MODIFIED:20190116T143344Z
UID:6388-1548759600-1548804600@www.loria.fr
SUMMARY:OLKi project presentation
DESCRIPTION:Christophe Cerisara (Synalp) will present OLKi Project on Tuesday\, January 29th at 11am in room A008. \nHis presentation will thus concern AI applied to natural language and knowledge\, open science and resource sharing through the Fediverse. \nMore about OLKi \n 
URL:https://www.loria.fr/event/olki-project-presentation/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190118T133000
DTEND;TZID=Europe/Paris:20190118T150000
DTSTAMP:20190109T091317Z
CREATED:20190109T091317Z
LAST-MODIFIED:20190109T091317Z
UID:6361-1547818200-1547823600@www.loria.fr
SUMMARY:Colloquium Loria : Félicien Vallet
DESCRIPTION:Félicien Vallet\, ingénieur au sein du service d’expertise technologique de la CNIL est le prochain invité du colloquium du Loria. \nRendez-vous vendredi 18 janvier à 13h30 dans l’amphithéâtre pour son exposé intitulé « Présentation de la CNIL et premiers enseignements à RGPD : M+8″. \n \nRésumé : \nEntré en application le 25 mai 2018\, le RGPD renforce les droits et devoirs en matière de protection des données. En offrant un cadre européen unifié\, il dote notamment les autorités de protection des données (comme la CNIL en France) de pouvoirs plus importants. Cette présentation vise à apporter des éléments d’information sur les fonctionnements\, rôles et missions de la CNIL ainsi que sur les nouveautés introduites par le RGPD. De premiers éléments d’appréciation « post-RGPD » seront également mis en avant. Enfin\, les aspects liés à la protection des données dans un cadre de recherche pourront être discutés à l’occasion de cette présentation qui se veut interactive et ouverte à tous. \n  \nMerci aux personnes extérieures au Loria de s’inscrire auprès de Marie Baron (marie.baron (at) loria.fr) avant mercredi 16 janvier.
URL:https://www.loria.fr/event/colloquium-loria-felicien-vallet/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190116T140000
DTEND;TZID=Europe/Paris:20190116T153000
DTSTAMP:20181213T135944Z
CREATED:20181213T135944Z
LAST-MODIFIED:20181213T135944Z
UID:6311-1547647200-1547652600@www.loria.fr
SUMMARY:D4 Seminar : Md Sahidullah
DESCRIPTION:Next TALC seminar will take place on Wednesday\, January 16th at 2pm in room A008. \nMd Sahidullah (Multispeech team)\, will give a presentation entitled « Speaker embeddings: from i-vector to x-vector and beyond ». \n \nAbstract:\nSpeaker recognition is the task of recognizing a human from his/her voice. The state-of-the-art speaker recognition technology uses a speaker embedding method for representing a speech utterance of arbitrary length in the form of a fixed-dimensional vector.  The recent advancements in deep neural network (DNN) research have enabled the\ndevelopment of robust and efficient speaker embedding techniques. In this talk\, I will first provide a brief overview of speaker recognition basics. It will be followed by the description of the conventional speaker embedding method popularly known as i-vector. Then I will present various attempts to develop speech signal representations with\nDNN-based discriminative training. I will explain the recently introduced x-vector embedding which showed promising speaker recognition performance. This talk will end with a discussion on potential future directions in the speaker embedding research including our ongoing work.
URL:https://www.loria.fr/event/d4-seminar-md-sahidullah/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190116T140000
DTEND;TZID=Europe/Paris:20190116T150000
DTSTAMP:20190110T165954Z
CREATED:20190110T165954Z
LAST-MODIFIED:20190110T165954Z
UID:6378-1547647200-1547650800@www.loria.fr
SUMMARY:Department 3 Seminar : Dr. Mauro Passacantando (University of Pisa)
DESCRIPTION:Next Department 3 seminar will take place on Wednesday\, January 16th at 2pm in room C005. \nDr. Mauro Passacantando (University of Pisa) will give a presentation entitled « Cooperative Infrastructure and Spectrum Sharing in Heterogeneous Mobile Networks ». \n \nAbstract: To accommodate the ever-growing traffic load and bandwidth demand generated by mobile users\, mobile network operators (MNOs) need to frequently invest in high spectral efficiency technologies and increase their hold of spectrum resources; MNOs have then to weigh between building individual networks or entering into network and spectrum sharing agreements. We address here the problem of radio access network and spectrum sharing in 4G mobile networks by focusing on a case when multiple MNOs plan to deploy small cell base stations in a geographical area in order to upgrade their existing network infrastructure. We propose two cooperative game models (with and without transferable utility) to address the proposed problem: for given network (user throughput\, MNO market\, and spectrum shares) and economic (coalition cost and mobile data pricing model) settings\, the proposed models output a cost division policy that guarantees coalition (sharing agreement) stability. \nCV: Mauro Passacantando received both M.Sc. and Ph.D. in Mathematics from the University of Pisa. He is currently Assistant Professor of Operations Research at the Department of Computer Science of the same university. He achieved the Italian National Habilitation for full professorship in Operations Research. His research activities are mainly focused on variational inequalities\, projected dynamical systems\, equilibrium problems\, and their applications. Most of his recent work focuses on theory and algorithms for Ky Fan inequalities and on game theoretic approaches to service provisioning problems in cloud and multi-cloud systems. He is author of one monograph and more than forty scientific papers published in international journals and volumes.
URL:https://www.loria.fr/event/department-3-seminar-dr-mauro-passacantando-university-of-pisa/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20190114T100000
DTEND;TZID=Europe/Paris:20190114T173000
DTSTAMP:20181219T133312Z
CREATED:20181219T133236Z
LAST-MODIFIED:20181219T133312Z
UID:6344-1547460000-1547487000@www.loria.fr
SUMMARY:Histoire et mémoire de l'informatique universitaire à Nancy (1950-2010)
DESCRIPTION:Lundi 14 janvier 2019 – 10:00 – 17:30\n\n\n\n\nUniversité de Lorraine (Nancy)\, Campus Lettres et Sciences Humaines\, Place Godefroy de Bouillon\, Salle G04\n\n\n********************************************************************************************************************************************** \nAprès Grenoble et Toulouse la Faculté des sciences de Nancy a été l’une des premières en France à développer des \n programmes de recherche et d’enseignement en « informatique ». À cette époque le terme n’existe pas encore dans la langue française et cette nouvelle discipline trouve d’abord ses origines dans l’analyse numérique et les mathématiques appliquées. \nAprès une année exploratoire\, tout commence par la mise en place par Jean Legras\, en 1958-59\, d’un cours de troisième cycle d’Analyse et de calcul numérique et par l’utilisation de calculateurs électroniques (IBM 604 puis 650). En 1961\, Jean Legras dirige la première thèse en « informatique » soutenue en France par Marion Créhange\, Structure du code de programmation. Suivront ensuite les travaux décisifs de Claude Pair sur le langage Algol et les théories de la programmation. \nSouvent mal considérée à ses débuts par les mathématiciens partisans des approches formelles des mathématiques – Nancy est à cette époque un des sièges du mouvement Bourbaki – cette nouvelle discipline s’installera cependant durablement dans le paysage universitaire local et national à travers différentes constructions institutionnelles qui parfois coexisteront au gré de regroupements et de changements de tutelle ou de direction : création d’un premier Centre de calcul en 1959\, fondation d’une équipe puis d’un Centre de Recherche en Informatique de Nancy (CRIN) en 1973-1976\, installation de l’Institut de Recherche en Informatique et Automatique (INRIA Lorraine) et de l’Institut de l’information scientifique et technique (INIST) dans les années 1980\, fondation du Centre de Recherche en Automatique de Nancy (CRAN) dans les mêmes années et du Laboratoire lorrain de recherche en informatique et ses applications (LORIA). \nL’histoire de l’informatique universitaire à Nancy dépasse largement le cadre strict de l’histoire de l’informatique ou même du « numérique ». Elle est une invitation à penser les processus de développement et d’installation d’une nouvelle discipline sur le temps long\, dans un contexte de concurrence scientifique et institutionnelle avec d’autres pôles universitaires. Elle appelle à penser le rôle de différents acteurs – professeurs\, ingénieurs\, techniciens\, étudiants – dans ce processus historique. Elle pose la question des processus institutionnels\, politiques et socio-économiques qui ont abouti à l’installation en Lorraine\, territoire fortement marqué par la désindustrialisation\, de structures de recherche et d’enseignement dédiées à cette discipline. Il s’agira donc de reconstituer l’écosystème informatique universitaire nancéien : non seulement ses lieux et ses modalités de formation et de recherche\, mais également ses relations avec les organismes de recherche nationaux (CNRS et INRIA)\, les collectivités locales\, les différents plans nationaux qui ont rythmé le développement de l’informatique en France et le monde économique et industriel. \nCette journée d’étude donnera la parole à ses principaux acteurs\, souvent eux-mêmes à l’origine de travaux historiques sur le sujet. Elle accueillera ainsi : Marion Créhange\, Jean-Claude Derniame\, Jean-Pierre Finance\, Marie-Christine Haton\, Jean-Paul Haton\, Bernard Legras\, Pierre Lescanne\, Claude Pair et Jean-Marie Pierrel. Elle sera ouverte par Pierre Éric Mounier-Kuhn\, historien de l’informatique\, auteur notamment de L’informatique en France\, de la Seconde Guerre Mondiale au Plan Calcul – L’émergence d’une science (2010). \nCette journée est ouverte au public\, sur inscription\, dans la limite des places disponibles. \nPour plus de détails écrire à Laurent Rollet. \n  \n\n\n\n\nProgramme: \n\n\n  \n10h00 : Ouverture de la journée \n10h15 : Pierre-Éric Mounier-Kuhn – « Histoire de l’informatique : historiographie et approches » \n10h35 : Bernard Legras – « Mon père Jean Legras (1914-2012)\, mathématicien lorrain\, promoteur de l’informatique en Lorraine » \n10h55 : Marion Créhange – « De 1957 à 1968 et même 1975 : une création clairvoyante et courageuse. Du défrichage au décollage » \n11h15 : Discussion \n11h30 : Claude Pair – « À tout CRIN : de la naissance à la maturité (1963-1976) » \n11h50 : Jean-Claude Derniame – « Du CRIN au LORIA (1981-1985) » \n12h10 : Discussion \n12h30 : Pause déjeuner \n14h00 : Jean-Pierre Finance – « De l’adolescence à l’âge adulte : dix ans du développement de la recherche en informatique à Nancy (1985-1994) » \n14h20 : Marie-Christine Haton – « 60 ans d’Informatique universitaire à Nancy\, une vue chronologique graphique » \n14h40 : Jean-Paul Haton – « L’intelligence artificielle à Nancy : une longue histoire » \n15h00 : Discussion \n15h15 : Pause \n15h30 : Pierre Lescanne – « Logique et informatique à Nancy » \n15h50 : Jean-Marie Pierrel – « Informatique et traitement numérique de la langue à Nancy : plus d’un demi-siècle d’histoire commune » \n16h10 : Discussion \n16h30 : Table ronde « Quelle histoire pour l’informatique universitaire nancéienne ? » \n\nPlus d’infos
URL:https://www.loria.fr/event/histoire-et-memoire-de-linformatique-universitaire-a-nancy-1950-2010/
CATEGORIES:Séminaire
ATTACH;FMTTYPE=image/jpeg:https://www.loria.fr/wp-content/uploads/2016/12/logo_loria_abrege.jpg
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20181220T133000
DTEND;TZID=Europe/Paris:20181220T150000
DTSTAMP:20181218T160836Z
CREATED:20181218T160836Z
LAST-MODIFIED:20181218T160836Z
UID:6342-1545312600-1545318000@www.loria.fr
SUMMARY:Séminaire SSL
DESCRIPTION:Next SSL Seminar will take place on Thursday\, December 20th at 1.30 pm in room C005. \nTayssir Touili (CNRS – LIPN) will give a presentation entitled « On static malware detection ». \n \nThe number of malware is growing extraordinarily fast. A malware may bring serious damage.\nThus\, it is crucial to have efficient up-to-date virus detectors. A robust malware detection technique needs  to check the behavior (not\nthe syntax) of the program without executing it. We show in this talk how using behavior  signatures allow to efficiently detect malwares in a completely static way. We applied  our techniques to detect several malwares. Our  tool was able to detect more than 800 malwares. Several of these malwares could not be detected by well-known anti-viruses such as Avira\, Avast\, Norton\, Kaspersky and McAfee
URL:https://www.loria.fr/event/seminaire-ssl/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;VALUE=DATE:20181214
DTEND;VALUE=DATE:20181215
DTSTAMP:20181206T080349Z
CREATED:20181119T114036Z
LAST-MODIFIED:20181206T080349Z
UID:6221-1544745600-1544831999@www.loria.fr
SUMMARY:Department 3 PhD day
DESCRIPTION:Departement 3 organizes its annual PhD Day on Friday\, December 14th in room C005\, to allow all the PhD students of the department to present their ongoing work and to share their experiences. \nSchedule \nProgram: \n\n\n\n9:00 – 9:05 Welcome (Ye-Qiong Song) \n9:05 – 10:30 session 1: Modeling\, simulation and experimental methods \nBéatrice Linot (20’ + 5’ Q&A)\nTrust in Computer-Supported crisis management information sharing \nThomas Paris (20’ + 5’ Q&A)\nComplex system modeling by composition \nAbdulqawi Saif (20’ + 5’ Q&A)\nContributions of experimental methods for I/O systems and testbed experiments \nJean-Baptiste Wiart (5’)\nDefinition of a domain specific language for the cosimulation of microgrid powered hydrogen \nThéo Docquier (5’)\nDesign\, modeling and co-simulation of real-time industrial IoT for smart grids \n10:30 – 10:50 coffee break \n10:50 – 12:00 session 2: Internet of Things\, SDN and security \nMingxiao Ma (20’ + 5’ Q&A)\nStudy of Synchronization Attacks on Distributed and Cooperative Control in Microgrid Systems \nGrégoire Denis (5’)\nMalicious attacks detection and resilience in cyber-physical systems through joint dynamic resource scheduling and synthesis of adaptive control laws \nAdrien Hemmer (5’)\nPredictive Security Monitoring for Large-Scale Internet-of-Things \nAbir LARABA (5’)\nData-Driven Intelligent Monitoring for Software-Defined Networks \nAhmad ABBOUD (5’)\nCompressed and Verifiable Filtering Rules in Software-defined Networking \n\n\n\n\n\n\nVirgile Dauge (20’ + 5’ Q&A)\nSystèmes Cyber-Physiques autonomes et communicants en milieux hostiles. Application à l’exploration par robots mobiles \n12:00 – 12:30 Invited session \nProf. Enrico Natalizio (25’ + 5’ Q&A)\n5G and UAVs: synergies to exploit for an Internet of Intelligent Things \n12:30 – 14:00 Buffet & Discussion \n14:00 – 15:15 session 3: Security \nHoang-Long Nguyen (20’ + 5’ Q&A)\nTransparency Approach using Blockchain to End to End Encryption (E2EE) \nNicolas Schnepf (20’ + 5’ Q&A)\nOrchestration and verification of security functions for smart environments \nVictorien Elvinger (20’ + 5’ Q&A)\nPrunable tamper-evident log in peer-to-peer systems \n15:15 – 15:30 coffee break \n15:30 – 17:10 session 4: Safety\, Performance and Optimization \nLouis Viard (20’ + 5’ Q&A)\nMonitor-Centric Mission Definition For Cyber-Physical Systems \nQuentin Laporte-Chabasse (20’ + 5’ Q&A)\nA topological characterisation of peer-to-peer inter-organisational collaboration \nBilal Messaoudi (20’ + 5’ Q&A)\nMultiple periods vehicle routing problems: a case study \nHoai-Le NGUYEN (20’ + 5’ Q&A)\nStudying group performance and behavior in collaborative editing
URL:https://www.loria.fr/event/department-3-phd-day-2/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20181207T100000
DTEND;TZID=Europe/Paris:20181207T113000
DTSTAMP:20181130T145539Z
CREATED:20181130T145532Z
LAST-MODIFIED:20181130T145539Z
UID:6283-1544176800-1544182200@www.loria.fr
SUMMARY:D4 Seminar : Philippe Muller (IRIT\, Toulouse)
DESCRIPTION:Next D4 Seminar will take place on Friday\, December 7th at 10am in room A008. \nPhilippe Muller (IRIT\, Toulouse) will give a presentation entitled « Sentential distributional semantics: Learning semantic sentence representations and their compositions ». \n \nAbstract: \n(Joint work with Damien Sileo et Tim van de Cruys) \nDistributional semantics aims at automatic representation of textual semantic content based on the observation of a large representative corpus. There is a large body of work on lexical distributional semantics\, based on the assumption that words appearing in similar contexts should have similar semantic representations. This popularized the representation of words as vectors in a semantic space.More recently\, a lot of effort in the NLP field has been devoted to building similar representations for sentences\, or even larger textual elements. \nThis raises several questions: how to build sentence representations from word representations in vector spaces\, preferably in a compositional manner\, and how to guide the representations so that they capture important semantic aspect at the sentence level?Arguably\, sequential compositional models such as recurrent neural network offer a simple composition at the lexical level that can be used in supervised settings to make accurate predictions in textual classification\, while building a representation of the sentential context in their internal state. This is however specific to each task\, and researchers have tried to find ways of building so-called « universal » sentence representations\, or more exactly transferable representations. In this perspective several settings have been proposed that evokes supervised distributional approaches at the word level\, with auxilliary tasks that could induce semantically relevant representations at the sentence level: for instance trying to predict if two sentences follow each other in a text\, or if one is a consequence of the other. These in turn must compose the two sentences in a way that allows for the learning of their relationships. Composition of representations is also important in all tasks that involve predicting a relation between a pair of textual elements: sentence similarity\, entailment\, discourse relations.\nThe compositions considered in NLP are often quite superficial\, and we will show more expressive compositions by taking inspiration from Statistical Relational Learning. Moreover we propose an unsupervised training task to induce sentence representations\, based on the prediction of discourse connections between sentences in a large corpus.
URL:https://www.loria.fr/event/d4-seminar-philippe-muller/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20181205T140000
DTEND;TZID=Europe/Paris:20181205T153000
DTSTAMP:20181130T150405Z
CREATED:20181130T150405Z
LAST-MODIFIED:20181130T150405Z
UID:6286-1544018400-1544023800@www.loria.fr
SUMMARY:TALC Seminar : Chloé Braud (Synalp)
DESCRIPTION:Next TALC Seminar will take place on Wednesday\, December 5th at 2pm in room A008. \nChloé Braud (Synalp) will give a presentation entitled « Transfer learning for discourse parsing ». \n \nAbstract:\nDiscourse structures describe the organization of documents in terms of discourse or rhetorical relations (such as « Explanation » or « Contrast ») linking clauses and sentences. Discourse analysis could be useful for various downstream applications\, such as automatic summarization\, question-answering or sentiment analysis. However\, the range of applications and the performance are still limited by the low scores of the existing discourse parsers and their focus on English. \nDiscourse parsing is known to be a hard task: It involves several complex and interacting factors\, touching upon all layers of linguistic analysis\, from syntax\, semantics up to pragmatics. Consequently\, also annotation is complex and time consuming\, and hence available annotated corpora are sparse and limited in size. \nIn this presentation\, I will present attempts to tackle these issues using transfer learning strategies. First\, I will describe experiments on identifying implicit discourse relations (i.e. lacking a discourse connective such as « but » or « because ») by transferring knowledge from the explicit examples to the implicit ones\, either by augmenting the size of the training set\, or by building a task-tailored representation of the words.I will then present two full discourse parsers. The first one involves a combination of several corpora annotated for different languages\, leading to improvements on English and to the first systems for Basque and Dutch developed without any training data. The second parser relies on multi-task learning to transfer information among several discourse related tasks.
URL:https://www.loria.fr/event/talc-seminar-chloe-braud-synalp/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;VALUE=DATE:20181205
DTEND;VALUE=DATE:20181207
DTSTAMP:20181127T161358Z
CREATED:20181127T161357Z
LAST-MODIFIED:20181127T161358Z
UID:6267-1543968000-1544140799@www.loria.fr
SUMMARY:QCM-BioChem days
DESCRIPTION:On the 5 and 6 of December\, the LORIA will host the meetings of the Mastodons Project: \nQCM-BioChem: Quality in Consensualizing and Mining biological and chemical datasets \nThe QCM-BioChem consortium is the result of the fusion of 3 Mastodons projects (QualiBioConsensus\, HyQual and DECADE) and that aims to address and tackle challenges arising from the large and wide variety of biological and chemical data as well as the different methodologies in knowledge discovery. The novelty of the approaches developed in QCM-BioChem relies on the consideration and explanation of quality criteria in terms of data\, and methods for data mining and data analysis. For further details:https://www.lri.fr/~cohen/QCM-BioChem.html \nThis meeting days will comprise 2 keynote presentations\,\n-Antti Kuusisto on December 5 at 14h00\, Room B013\n-Jérôme Lang on December 6 at 14h00\, Room C005\nfollowed by several contributed presentations. \nProgram of December 5 (room B013): \n14h00-15h00: Antti Kuusisto (Tampere University of Technology)\nTitle: Computational Logics – A Theoretical Perspective \n15h15-15h45: Arnaud Soulet (LI\, Université de Tours)\nTitle: Representativeness of knowledge bases with the generalized Benford’s Law \n15h45-16h15: Coffee break \n16h15-16h45: Tatiana Makahlova (LORIA\, Université de Lorraine)\nTitle: The Application of MDL in the Mining of Numerical Data \n17h00-17h45: Sylvie Hamel (Université de Montréal)\nTitle: Space reduction techniques for the median of permutations problem \nProgram of December 6 (room C005): \n10h00-10h45: Henry Soldano (LIPN\, Université Paris-Nord)\nTitle: Bi-pattern mining of attributed two-mode and directed networks \n10h45-11h00: Coffee break \n11h00-11h30: Lamine Diop (LI\, Université de Tours)\nTitle: Sequential Pattern Sampling with Norm Constraints \n11h45-12h15: Nyoman Juniarta (LORIA\, Université de Lorraine)\nTitle: Application of Biclustering to the discovery of constant and gradual Patterns \n12h30-14h00: Lunch break \n14h00-15h00: Jérôme Lang (LAMSADE\, Université Paris-Dauphine)\nTitle: From social choice to preference learning \n15h00-15h30: Coffee break \n15h30-16h00: Guilherme Alves (LORIA\, Université de Lorraine)\nTitle: A framework for online clustering based on evolving semi-supervision \n16h15-16h45: Kevin Dalleau (LORIA\, Université de Lorraine)\nTitle: Tackling (some) preprocessing issues using unsupervised extremely randomized trees \n======================================= \nThe complete program together with the abstracts\, of all presentations of this event can be found at:\nhttps://malotec.loria.fr/journees-mastodons-qcm-biochem
URL:https://www.loria.fr/event/qcm-biochem-days/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20181129T133000
DTEND;TZID=Europe/Paris:20181129T150000
DTSTAMP:20181114T161613Z
CREATED:20181114T161613Z
LAST-MODIFIED:20181114T161613Z
UID:6211-1543498200-1543503600@www.loria.fr
SUMMARY:MFX Seminar : Jean-Baptiste Labrune
DESCRIPTION:Jean-Baptiste Labrune will give a seminar on Thursday\, November 29th at 1:30 pm in the Amphitheater. \nHis presentation is entitled « Radical Design: augmented environments for building dynamic objects with active matter ». \n \n\nAbstract :\n\nProgrammable matter\, also called active matter\, is a class of physical and biological materials that allow new kind of affordances and interactions (Tibbits\, Active Matter\, MIT Press\, 2017). Static entities become dynamic and can shapeshift\, turn elastic or solid on demand. Active matter is still a prototype in laboratories\, hence not easily available nor simple to use for non-experts.\nHowever\, like VLSI opened the field for advanced software engineering\, technologies such as 3D / 4D printing and bio/nano-assemblers will probably popularize them soon. In this new context\, how will we program but also design with these new materials? What will be the appropriate tools and interaction paradigms to build this world that Ivan Sutherland defined as « ultimate » ?\nOn the basis of my research as PhD at INRIA and postdoc at the MIT Medialab\, i will show some initial answers to these questions. I will detail in particular how the vision of Radical Atoms (Ishii\, Labrune & al 2012) proposes a new kind of paradigm called Radical Design\, connecting biology\, material science\, computational modelling of microstructures\, HCI and design.\n\n\n—\n\n\nJean-Baptiste Labrune is a designer & researcher\, specialized in the study of creative places and processes. His researches focus on the notion of « Exaptation »\, the way in which users of technology reconfigure and hack it\, producing original and unexpected functions and uses. He completed his Phd at INRIA and postdoc at MIT\, then became researcher at Bell Labs and interaction design professor at ENSAD (Arts Décos School). He organizes and lead many « hybrid » workshops in art & sciences places in France (Arts Décos\, Beaux-Arts\, Palais de Tokyo\, Mains d’Oeuvres) & internationally (Mediamatic\, Interaction Design Institute Ivréa\, IMAL\, Hangar\, Hyperwerk\, Akademie Schloss Solitude\, MIT Medialab).\n\n\nhttp://radicaldesign.eu
URL:https://www.loria.fr/event/mfx-seminar-jean-baptiste-labrune/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20181127T133000
DTEND;TZID=Europe/Paris:20181127T150000
DTSTAMP:20181119T095640Z
CREATED:20181113T160453Z
LAST-MODIFIED:20181119T095640Z
UID:6199-1543325400-1543330800@www.loria.fr
SUMMARY:Colloquium Loria : Claire Mathieu
DESCRIPTION:Claire Mathieu is the next speaker of Loria’s colloquium. \nShe will give a presentation entitled « Stable Matching in Practice » on Tuesday\, November 27th at 1:30 pm in the Amphitheater. \nAbstract \nStable matching methods\, based on the algorithm designed by Gale and Shapley\, are used around the world in many applications such as college admissions. Several criteria measure the quality of the result: number of students assigned; rank of the college assigned to the applicant in their preference list; robustness; running time; etc. \nAfter reviewing properties of the algorithm in the pure\, ideal setting\, we present issues arising in practice. The input data is uncertain and evolves with time\, so a one-shot algorithm does not suffice. It is not feasible for admission committees to meet continuously\, so the process cannot be fully dynamic. To reconcile those competing constraints\, a hybrid implementation proceeding partly online on the student side was recently proposed for college admissions in France. The system also incorporates side constraints on joint assignment to schools and to dorms. \nAbout the Speaker\nClaire Mathieu does research on the design and analysis of algorithms\, with a focus on approximation algorithms\, particularly approximation schemes for NP-hard problems. A former student of Ecole normale supérieure\, she received a PhD in Computer Science in 1988 at Paris-Sud University. She has held research and faculty positions at CNRS\, Paris-Sud University\, Ecole Polytechnique\, Brown University\, and Collège de France. She is currently a CNRS research director in Paris\, France. \n  \nTo attend the colloquium\, people from outside the laboratory may send an email to marie.baron (at) loria.fr before Friday\, November 23rd.
URL:https://www.loria.fr/event/colloquium-loria-claire-mathieu/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20181121T140000
DTEND;TZID=Europe/Paris:20181121T160000
DTSTAMP:20181003T065520Z
CREATED:20181003T065520Z
LAST-MODIFIED:20181003T065520Z
UID:6035-1542808800-1542816000@www.loria.fr
SUMMARY:TALC Seminar : Emmanuel Dupoux
DESCRIPTION:Next TALC Seminar will take place on Wednesday\, November 21\, at 2pm in room A008. \nEmmanuel Dupoux (EHESS\, Laboratoire de Sciences Cognitives et Psycholinguistique) will give a talk entitled « Towards developmental AI ». \nAbstract:\nEven though current machine learning techniques yield systems that achieve parity with humans on several high level tasks\, the learning algorithms themselves are orders of magnitude less data efficient than those used by humans\, as evidenced by the speed and resilience with which infants learn language and common sense. I review some of our recent attempts to reverse engineer such abilities in the area of unsupervised or weakly supervised learning of speech representations and speech terms\, and the learning the laws of intuitive physics by observation of videos. I argue that a triple effort in data collection\, algorithm development and fine grained human/machine comparisons is needed to uncover these developmental algorithms.
URL:https://www.loria.fr/event/talc-seminar-emmanuel-dupoux/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;VALUE=DATE:20181120
DTEND;VALUE=DATE:20181121
DTSTAMP:20181119T115026Z
CREATED:20181119T115026Z
LAST-MODIFIED:20181119T115026Z
UID:6223-1542672000-1542758399@www.loria.fr
SUMMARY:Department 2 PhD Day
DESCRIPTION:Department 2 PhD Day will take place on Tuesday\, November 20th. \nProgram: \n  \n9h15-9h30 Accueil (croissants) \n9h30-10h30 Première session \n\nJoseph Lallemand – Voting: You Can’t Have Privacy without Individual Verifiability\nHans-Jorg Schurr – Towards really neat proofs for SMT\n\n10h30-11h00 Pause café \n11h00-12h30 Deuxième session \n\nNicolas Schnepf – Orchestration and verification of security functions for smart environments\nSylvain Cecchetto – BOA : a CFG builder (by Basic blOck Analysis) based on system state prediction\nPierre Lermusiaux – Analysing algebraic type transformations with Pass Rewriting\n\n12h30-14h00 Repas (Pizzas) \n14h-15h30 Troisième session \n\nCharlie Jacomme – Symbolic methods applied to the automation of computational proofs\nRenaud Vilmart – A Simplified ZX-Calculus for Universal Quantum Computing\nDaniel El Ouraoui – Machine learning for instance selection in SMT solving\n\n15h30-16h Café et croissants \n16h-17h Dernière session \n\nMargaux Duroeulx – Satisfiability techniques for assessing systems reliability\nItsaka Rakotonirina – Symmetries of equivalence in security protocols\n\n\n \nSylvain Cecchetto – BOA : a CFG builder (by Basic blOck Analysis) based on system state prediction\nThis research is targeting people who want to study a binary for which the do not have the source code. These people include malware analysts who want to be able to understand the behavior and/or the dangerousness of an unknown program. We can also think about reversers who might want to be able to extract and recover the code of a specific part of a binary\, or\, also to analyse a program used in a constrained environment like an airplane\, in order to evaluate the risks that it crash or reach an unwanted state. Overall\, this research can be useful in domains like malware detection\, reverse engineering and system protections. In order to analyse the behavior of a program for which we do not have the source code (typically in a malware analysis case)\, it is almost mandatory for the analyst to disassemble its binary code\, that is to list all program instructions. However\, as Schwarz\, Debray ans Andrews explained\, the disassembly step is undecidable due to dynamic jumps instructions. Unfortunately\, the problem remains\, and partial solutions are based on heuristics. Two properties are desired during this step\, (i) the completeness that aims to recover all instructions that the binary contains without missing and (ii) the correction that aims to only disassemble instructions that can be executed. Furthermore\, we have to deal with obfuscated binaries. Obfuscation techniques are designed to protect code/program against human and automated analysis. The obfuscation can be legitimate to protect intellectual property or doubtful in the case of a malware because the author want to slow down the analyst. In 2003\, Linn and Debray expose some obfuscated techniques in order to make loose any standard disassembly tool. On another side\, Kruegel\, Robertson\, Valeur and Vigna in 2004 propose to add some improvement on existing disassembly algorithms in order to be able to disassemble obfuscated binaries like the ones build with the Linn and Debray techniques. Finally we are facing the game of cat and mouse between\, in the first side\, always better obfuscated techniques and in other side\, disassembly tools that have to be more and more robust. Usually\, the disassembly step is followed by the control flow graph reconstruction. This step aims to construct an oriented graph from the basic blocks (previously obtained during disassembly) in which the links describe the different paths that the binary can take during an execution. This construction is undecidable in general since it involves variable values during execution. Furthermore\, as we said\, obfuscation techniques like self-code-modification\, call stack tampering or opaque predicate can make difficult both\, disassembly and control flow graph constructions steps. \nIn order to improve the disassembly and control flow graph constructions steps some works have be done based on symbolic execution techniques like it is shown by Yadegari and Debray in 2015. As Schwartz\, Avgerinos and Brumley said in 2010\, this method allows to reason about the behavior of a program on many different inputs at one time by building a logical formula that represents a program execution. In this paper we propose a new approach called BOA (for Basic blOck Analysis) to solve (at least partially) the disassembly and control flow graph construction problems. This method is based on symbolic execution at a basic block level. This technique aims to compute a partial system state as pre and post conditions of BB. Then\, from this partial system state we can compute dynamic jump/call instructions target addresses\, detect call stack tampering\, opaque predicate or self-code-modifications. \n\n \nMargaux Duroeulx – Satisfiability techniques for assessing systems reliability\nFault trees (FTs) or reliability block diagrams are commonly used representations for the reliability assessment of systems. In our work\, we aim at using satisfiability (SAT) techniques as a building block in this analysis\, based on the computation of minimal cut sets or tie sets from the system’s structure function. We previously introduced the use of SAT techniques for computing tie sets of systems with a structure function that depends only on the working status of components. We now extend the approach to systems where the order of failures matters\, and that can be represented by a dynamic fault tree (DFT). \nBeyond the gates available in an ordinary FT (such as conjunction\, disjunction or k out of n)\, a DFT may contain three types of dynamic gates indicating priority conjunction\, functional dependency\, and spare components. The corresponding structure function depends not only on the working status of components but also on the temporal order between component failures\, represented by additional propositional variables in the SAT encoding. The appropriate generalization of a tie set is then a Tie Set with Sequence (TSS). Whereas an ordinary tie set only reflects the set of components that are functioning correctly\, a TSS also records the order in which component failures occurred. Standard tie and cut sets can be organized in a Hasse diagram that reflects the complete partial order between these configurations due to component failures\, and the Hasse diagram provides a convenient basis for computing the reliability function. We show that this representation extends to tie and cut set sequences. \nOur approach generates the structure function from a DFT\, uses a SAT solver for efficiently computing the minimal TSSs\, and obtains the reliability polynomial based on the positions of the TSSs in the Hasse diagram. According to the type of analysis the user is interested in\, our method can therefore provide qualitative results in the form of tie set sequences or quantitative results by computing the probability of system failure after a given period of time. \n\n \nDaniel El Ouraoui – Machine learning for instance selection in SMT solving\nSatisfiability modulo theories (SMT) solvers have increased their capabilities to solve many different problems with the development of ever more effective decision procedures. Hence state of the art SMT solvers have now established themselves as one of the best solutions to tackle ground quantifier-free first order problems. Nevertheless\, automated theorem provers (ATP) based on superposition techniques remain incredibly better to handle quantified formulas. One reason is that they are fully understanding the semantics of quantifiers\, whereas SMT solvers try to heuristically find a set of instances which could lead to solve the grounded problem. Often\, these techniques produce a very large number of instances\, many of them being useless. Generally\, only 10% of the instances are useful. This often explains why SMT solvers fail to prove simple first order problems. We suggest to apply state of the art machine learning techniques as classifiers for instances on top of the process. We show that such techniques can indeed lead to significantly smaller numbers of generated instances\, and even more\, can be used to solve new problems that could not be solved before. \n\n \nCharlie Jacomme – Symbolic methods applied to the automation of computational proofs\nAfter having introduced two classical models used in the formal verification of security protocols – the Symbolic model and the Computational model – we will show how the former may be used to improve proofs automation in the latter. \n\n \nJoseph Lallemand – Voting: You Can’t Have Privacy without Individual Verifiability\nElectronic voting typically aims at two main security goals: vote privacy and verifiability. These two goals are often seen as antagonistic and some national agencies even impose a hierarchy between them: first privacy\, and then verifiability as an additional feature. Verifiability typically includes individual verifiability (a voter can check that her ballot is counted); universal verifiability (anyone can check that the result corresponds to the published ballots); and eligibility verifiability (only legitimate voters may vote). We show that actually\, privacy implies individual verifiability. In other words\, systems without individual verifiability cannot achieve privacy (under the same trust assumptions). To demonstrate the generality of our result\, we show this implication in two different settings\, namely cryptographic and symbolic models\, for standard notions of privacy and individual verifiability. Our findings also highlight limitations in existing privacy definitions in cryptographic settings. \n\n \nPierre Lermusiaux – Analysing algebraic type transformations with Pass Rewriting\nProgram transformation is a common practice in computer science\, and its many applications can have a range of different objectives. For example\, a program written in an original high level language could be either translated into machine code for execution purposes\, or towards a language suitable for formal verification. In all cases\, the transformation is often realized in several phases\, called passes\, each performing a smaller transformation\, thus going through a number of intermediate languages. \nWhile languages often have a lot of different constructions\, only a few are generally concerned for each one of these passes. Some languages\, such as Tom and Stratego\, use rewriting strategies to effectively implement traversal operation so as to not have to deal with unchanged constructions. Another interesting solution\, with a similar pattern matching approach\, is proposed by the Nanopass framework: using autogeneration techniques on these same trivial cases. In order to ensure the correctness of the overall transformation\, w.r.t. the target language\, each pass should be able to guarantee that the intermediate program is well-formed in the context of its corresponding language. However\, neither mentioned solution manages to satisfyingly provide such guarantees. \nTherefore\, we propose a new formalism for transformation techniques\, based on pattern matching and rewriting\, with an expressive power comparable to above solutions. Moreover\, it will enable statically performing type analysis on the transformation\, in order to improve on their limitation. The interest of this formalism will be largely based on this type analysis\, which would allow the definition of transformations in a general context while still providing syntactical guarantees. \n\n \nItsaka Rakotonirina – Symmetries of equivalence in security protocols\nSecurity protocols are distributed programs specifying how some agents can exchange some messages remotely\, through an untrusted network. These protocols are expected to offer some privacy guarantees against dishonnest parties controlling the communication network\, but their analysis has proved tedious and error-prone when performed by hand. Efficient automated tools exist to support flaw detection\, e.g. the recent DEEPSEC prover\, but still have some shortcomings in terms of scalability (the underlying theoretical problem exhibiting a high complexity). In this work we present techniques exploiting the properties that typical systems have\, e.g. internal symmetries\, to reduce the practical cost of performing security analyses and therefore improve on the scope automated analysers. An implementation of these techniques is in progress. \n\n \nNicolas Schnepf – Orchestration and verification of security functions for smart environments\nSecurity threats against smart environments are exponentially growing for several years due to lack of market preventive methods. A solution proposed by researchers consists in chaining security functions for dynamically protecting those devices: in particular\, those chains would benefit of the programmability provided by software defined networks (SDN) for automating their deployment and their adjustment. Nevertheless\, the multiplication and the complexity of such chains of security functions increase the risk of introducing misconfigurations in network policies: because of this complexity\, the validation of such chains require the use of formal methods for guarantying their correctness before their deployment. \nThe goal of this PhD is to design a framework for the orchestration and the verification of chains of security functions. In our previous work we already designed an approach for the validation of security policy called synaptic: this framework relies on formal methods for validating the correctness of SDN policies. Complementary to this work we proposed an approach for automatically profiling android applications in order to identify their security requirements. The remaining part of our work will consist in designing an approach for automatically generating or selecting chains of security functions corresponding to the applications running on a device. \n\n \nHans-Jorg Schurr – Towards really neat proofs for SMT\nveriT is a state of the art Satisfiability Modulo Theories (SMT) solver. Given an unsatisfiable input problem\, veriT can output a certificate – a proof – of its unsatisifiability. These proofs serve at least two purposes. On the one hand\, they can be checked by an external proof checker\, relieving the SMT solver from the responsibility of being trustworthy\, on the other hand they can be used by client systems for further processing. In the past the proofs generated by veriT have been partially reconstructed in the interactive theorem prover Isabelle/HOL. We discuss ongoing work on improving SMT proof production. While work is constrained by the architecture of the SMT solver\, the requirements on the format are given by typical use cases. Overall\, our goal is to ensure veriT produces proofs which are easy to check\, while maintaining performance. \n\n \nRenaud Vilmart – A Simplified ZX-Calculus for Universal Quantum Computing\nThe ZX-Calculus is a powerful graphical language for reasoning about quantum processes and quantum computing. It has various applications\, ranging from foundations of quantum mechanics and models of quantum computation\, to compilation of quantum circuits\, quantum error correction\, … Complete axiomatisations have recently been provided for the first approximately universal fragment of the language\, and right after that for the whole language. These axiomatisations\, even though complete\, suffered one major drawback: some of their axioms involve a lot of nodes\, and consequently they are hard to use\, and their interpretations are far-fetched\, if not non-existent. We provide a simplified axiomatisation for the general ZX-Calculus\, giving the calculus one of its main initial features back: intuitiveness.
URL:https://www.loria.fr/event/department-2-phd-day-2/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20181115T140000
DTEND;TZID=Europe/Paris:20181116T000000
DTSTAMP:20181017T130129Z
CREATED:20181017T130129Z
LAST-MODIFIED:20181017T130129Z
UID:6115-1542290400-1542326400@www.loria.fr
SUMMARY:Colloque Cathy Dufour 2018 : Intelligences Artificielles
DESCRIPTION:Le colloque Cathy Dufour – Intelligences Artificielles aura lieu dans l’amphi 5 de la Faculté des Sciences et Technologies de Vandoeuvre-les-Nancy le 15 et 16 novembre. \nAu programme : \nJeudi 15 novembre \n14h – 14h 35 : Nazim Fates (Inria Nancy)\nIntelligence artificielle : vers l’« ordination universelle » ? \n14h 40 – 15h 15 : Maxime Amblard (Université de Lorraine)\nCalculer sur la langue mais qu’y comprendre ? \n15h 15 – 15h 45 : Discussion \n16h 15 – 16h 55 : Manuel Rebuschi & Marion Renauld (Université de Lorraine)\nInteragir avec une machine ou faire-semblant ? \n16h 55 – 17h 30 : Irène Marcovici (Université de Lorraine)\nAutomates cellulaires et phénomènes d’auto-organisation : le rôle de l’aléa \n17h 35 – 18h 10 : Alain Dutech (Inria Nancy)\n« Deep Reinforcement Learning » : des fois ça marche\, souvent ça marche pas ! \n18h 10 – … : Discussion \nVendredi 16 novembre \n9h 15– 9h 50 : Marianne Clausel (Université de Lorraine)\nModélisation probabiliste et analyse de données textuelles : les approches de type topic modeling \n9h 55 – 10h 30 : Thomas Boraud (Université de Bordeaux) \n11h – 11h 35 : Frédéric Alexandre (Inria Bordeaux)\nL’Intelligence Artificielle apprend-elle de ses erreurs ? \n11h 35 – … : Discussion
URL:https://www.loria.fr/event/colloque-cathy-dufour-2018-intelligences-artificielles/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20181115T133000
DTEND;TZID=Europe/Paris:20181115T150000
DTSTAMP:20181008T081839Z
CREATED:20181008T081838Z
LAST-MODIFIED:20181008T081839Z
UID:6088-1542288600-1542294000@www.loria.fr
SUMMARY:SSL Seminar : Corinna Schmitt
DESCRIPTION:Next SSL Seminar will take place on Thursday\, November 15th at 1.30 pm in room C005. \nCorinna Schmitt (Universität der Bundeswehr München) will give a presentation entitled « Authentication in IoT Networks ». \nToday over 35 billion devices are connected with each other building the Internet of Things (IoT). The device diversity ranges from constrained devices (e.g.\, sensor\, Smartwatches) over Tables and Smartphone to resource-rich devices like notebooks and servers. In parallel the stack in IoT shows also diversity and includes usage of many standards and third-party services at the same time from collection point to the application. Manifold data is collected all the time and the users have less knowledge about it\, but their awareness of misuse rises.\nBased on this quite complex situation\, authentication in IoT networks is important. But which authentication are we speaking of here? In general two opportunities exist: (1) authentication within the deployed network (e.g.\, between the devices using encryption and handshakes) and (2) authentication from the user side controlling the access. In this talk the focus is placed on the second opportunity\, namely user authentication. Most techniques are smartcard based\, but the Web-based approach developed within SecureWSN is different: It is based on credentials and automatically handled requests without involvement of third-parties giving data owner full control of access. Thus\, first features of the GDPR strengthening ownership are included in SecureWSN. \nBio: Corinna Schmitt holds a Diploma in Bioinformatics (Dipl. Informatik (Bioinformatik)) from the Eberhard-Karls University of Tübingen (Germany) and a Doctor in Computer Science (Dr. rer. nat.) from the Technische Universität München (Germany). She established an efficient data transmission protocol – called TinyIPFIX – with additional features for aggregation\, compression\, and secure transmission\, complementing it with an user-friendly and flexible GUI (CoMaDa).\nFrom spring 2013 to May 2018 she was employed at the University of Zurich (Switzerland) as « Head of Mobile and Trusted Communications » at the Communication Systems Group (CSG) of Prof. Dr. B. Stiller. Her focus was on constrained networks\, security and privacy issues\, as well as on Internet of Things related issues. After several years of visiting status at the goup of Prof. Dr. Gabi Dreo-Rodosek at the Universität der Bundeswehr München (Germany) she joint the affiliated Research Institute CODE as researcher and laboratory supervisor. Her research focuses is the same as at the CSG-Group with expands to the application area of military communication and Smart City.\nHer work is documented in more than 30 publications\, including 8 book chapters\, the RFC 8272 on « TinyIPFIX for Smart Meters in Constrained Networks »\, and the ITU-T recommendation Y.3013 on « Socio-economic Assessment of Future Networks by Tussle Analysis ». She contributes / contributed to several EU projects (e.g.\, CONCORDIA\, AutHoNe\, SmartenIT\, FLAMINGO\, symbIoTe) and different standardization organizations (IETF\, ITU\, ASUT) until now and continues with these activities and recruits research funds continuously. She is active in ACM and IEEE as TCP member\, as well as reviewer for several journals and organizer of conferences.
URL:https://www.loria.fr/event/ssl-seminar-corinna-schmitt/
CATEGORIES:Séminaire
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Europe/Paris:20181114T140000
DTEND;TZID=Europe/Paris:20181114T153000
DTSTAMP:20181022T112723Z
CREATED:20181022T112723Z
LAST-MODIFIED:20181022T112723Z
UID:6130-1542204000-1542209400@www.loria.fr
SUMMARY:TALC Seminar : Antoine Deleforge
DESCRIPTION:Next TALC seminar will take place on Wednesday\, November 14th at 2pm in room C005. Antoine Deleforge (Multispeech) will give a presentation entitled « Audio signal processing with a little help from echoes ». \nAbstract:\nWhen a sound wave propagates from a point source through a medium and is reflected on surfaces before reaching microphones\, the measured signals\nconsist of mixtures of the direct path signal with delayed and attenuated copies of itself. This acoustical phenomenon is referred to as echoes\, or reverberation\, and is generally considered as a nuisance in audio signal processing. After introducing some basic signal processing and acoustic background\, this seminar will present recent works showing how acoustic echoes can be blindly estimated from audio recordings\, and how the knowledge of such echoes can actually help some audio signal processing tasks such as beamforming\, source separation or sound source localization. \nNext seminars: http://talc.loria.fr/
URL:https://www.loria.fr/event/talc-seminar-antoine-deleforge/
CATEGORIES:Séminaire
END:VEVENT
END:VCALENDAR