Enseignant Chercheur



Contact information:


+33 2 29 00 14 27

    Personal Page

    I am a member of the Computer Science Departement of IMT Atlantique, as well as of the research team DECIDE, belonging to the CNRS laboratory Lab-STICC. My research topics are natural language processing, text mining and grapholinguistics. I'm teaching graph theory, logic, probabilistic graphical models and computational linguistics.


    Born in Athens (Greece) in 1962, I graduated from the Lycée Léonin High School (Athens) in 1979. I obtained my bachelor (DEUG A, Licence de mathématiques, Maîtrise de mathématiques pures and partly Maîtrise de mathématiques appliquées) and my master (DEA de mathématiques pures) from University of Lille I (USTL). I graduated (thèse de doctorat ès sciences mathématiques) in 1990 at the same university, in the field of Algebraic Topology, under the supervision of Prof. Daniel Tanré.

    During my Ph.D. and until 1992, I worked as assistant associate (assistant associé and later ATER) at University of Lille I, giving classes in Algebra, Differential Calculus and Analysis.

    After 1990 I became active in the field of digital typography and internationalization. In 1993/94 I gave lectures (chargé d'enseignement) at the Institute of Oriental Languages and Civilisations (INALCO) in Paris as an external member of the CERTAL (Centre d'études et de recherches sur le traitement automatique des langues). In 1994 my wife and I started the company Atelier Fluxus Virus, specialized in the development of digital typography tools and systems. In 2003 I wrote a 1,000+ page book on digital typography and internationalization (Fontes et codages) for O'Reilly France, which was subsequently translated into English and published by O'Reilly US (Fonts & Encodings, From Advanced Typography to Unicode and everything in between, 2007).

    In 2001 I entered IMT Atlantique (at that time: ENST Bretagne) as a full Professor (directeur d'études) at the Computer Science Departement. I continued research on digital typography and electronic documents, progressively orienting myself to the area of natural language processing and text mining.

    In 2010 I became a full member of the DECIDE research group (Data Mining and Decision Theory) of the Lab-STICC CNRS laboratory.

    Ever since 2011 I am in charge of the DS (Data Science) Master of Science of IMT Atlantique. Some of my classes are specific to this master (natural language processing, statistics) and others (probabilistic graphical models, graph theory, logic) are open also to other students of IMT Atlantique.

    In 2020 I obtained my “Habilitation à diriger des recherches” from Université de Bretagne Occidentale.

    Here are my identifiers: ORCID (0000-0003-1443-6115), HAL, IdRef, WikidataGoogle Scholar, Semantic Scholar, dblp, Mathematics Genealogy Project,

    My Erdős Number is 3 (Erdős—Diaconis—Rowley—me).

    In 2018, 2020, and 2022 I organized three editions of the international conference Grapholinguistics in the 21st Century. All talks have been recorded and can be viewed on the related Web sites (2018, 2020, 2022). The Proceedings have been published by Fluxus Editions


    Cheque of one hexadecimal dollar by Donald Knuth

    Cheque of 1 hexadecimal dollar emitted by Donald E. Knuth on March 20, 2018

    Knuth chèque

    Cheque of 1 hexadecimal dollar emitted by Donald E. Knuth on May 5, 2021


    1. Yannis Haralambous, Formation doctorale/LIESSE LaTeX. En supplément, page en arabe et chinois.
    2. Yannis Haralambous, Poster presented at the 12th International Workshop of the Association for Written Language and Literacy (Cambridge, March 26-28, 2019): Phonocentrism in Greece: Side Effects of Two Centuries of Diglossia.
    3. Yannis Haralambous, Introduction au grec moderne pour la classe de grec du Lycée Sainte-Anne.
    4. David I. Masson, Hausfreund von Vorgeſtern (1968), übersetzt von Horst Pukallus, gesetzt in Schönberger Fraktur und HW Caslon.
    5. L'interpréteur de logique combinatoire de John Tromp :

      L'archive cl.tar.gz provient de la page dédiée à la logique combinatoire de John Tromp (Tromp avec un `o' !!, un adepte du jeu de Go), pour interprêter des programmes en logique combinatoire il faut le décompresser, compiler :


      et ensuite l'exécuter :

      java CL

      Sur la ligne de commande on peut écrire des expressions qui séront réduites (par exemple : en écrivant KKxyz on obtient la réponse :

      of size 1
      head reduces in 0 steps to y of size 1

      On peut également utiliser une fonction anonyme f, on obtient alors le combinateur complexe qui a le même comportement, par exemple si on demande fxyzt=t (le projecteur π44), on obtient

      defines f as K(SK)
      of size 3

    6. Yannis Haralambous, Transparents de la présentation au colloque HUMANUM, Cluj-Napoca, 18 octobre 2018.
    7. Yannis Haralambous & Philippe Lenca, Response to the “My Tailor is Rich” Challenge, May 28, 2018.
    8. Yannis Haralambous, La graphématique, de Saussure à Unicode, et au-delà..., Conférence dans le cadre du séminaire du Master Sciences du langage, Université Grenoble-Alpes, le 21 septembre 2018.
    9. François Sébastianoff, La poligrafi, une écriture pour tous les usages du français, 1990.
    10. Corpus de mails ENRON pour formation LIESSE «Traitement automatique de la langue».
    11. Small manual for the CCG LaTeX style file by B. Hoffman, M. Steedman, and J. Baldridge
    12. Annexe C de l'article «Exemples de requêtes SPARQL associées aux patrons de questions» de l'article «Un système de questions-réponses automatiques dans le domaine légal : le cas des réglementations maritimes», TAL 58 (2) 47–72, 2018.
    13. PDF file 1 of graphs and PDF file 2 of graphs of the paper An Attempt to Decrypt Pages 269-271 of Jonathan Safran Foer's Extremely Loud & Incredibly Close, published on viXra.
    14. Graphemic and graphetic methods for gender-neutral writing, Invited conference at the Technische Universität Braunschweig, January 14th, 2021. Youtube link.
    15. Translation of the xeCJK manual into English by DeepL. This is a translation of the manual of the excellent LaTeX package xeCJK into English. The original, on the ctan Web site, is only in Chinese. This is an automatic translation so it is not always perfect but it is better than Chinese for those who don't speak this language.
    16. A Gephi-Generated graph of the largest connected component of WordNet 3.1's nouns (74,374 synsets, 75,834 hyperonymic relations).