Dan Tufis

Institutul de Cercetari pentru Inteligenta Artificiala al Academiei Romane, Bucuresti, .

E-mail: trimite un mesaj.

Pagina web a instituţiei: http://www.racai.ro
Pagina web personala: http://www.racai.ro/~tufis

Nascut(a) in: 1954

Interese: lingvistica corpusului, prelucrarea limbajului natural, traducere automata, tagging, wordnet, ontologii lexicale, aliniere lexicala, dezambiguizare semantica, reprezentarea cunostintelor

Detalii:
Domenii de cercetare: lingvistica corpusului, instruirea inteligenta, asistata de calculator in invatarea limbilor straine, invatarea automata, traducere automata, generarea limbajului natural, reprezentarea cunostintelor.
Prof. Tufis este principalul autor al primelor sisteme de prelucrare a limbii romane SDLR, IURES, PC-IURES, PARADIGM si de asemenea principalul autor al mediului de programare TC-LISP. Prof. Tufis si colegii sai au dezvoltat primele corpusuri standardizate pentru limba romana (TEI/CES, adnotate morfo-lexical si validate pana la nivelul cuvantului) si a resurselor lingvistice asociate (wordnetul romanesc, DEX-dictionarul explicativ-in format XML conform satndardului CONCEDE, morfologia paradigmatica a limbii romane, descrierea formala a segmentarii lexicale, reguli de interpretare a cuvintelor necunoscute, sistemul de etichetare morfo-lexicala pentru limba romana, conform cu standardul MULTEXT-EAST, modele statistice pentru limba romana.
Prof. Dan Tufis este autorul modelului de etichetare morfo-lexicala etichetata pentru dezambiguizarea statistica, precum si al modelului reificat de aliniere lexicala.
Prof. Tufis a sustinut prelegeri invitate la diferite universitati si institute de cercetare din Europa, Asia, si America de nord (USA si Canada): Belgia (Vrije University of Brussels), Bulgaria (Bugarian Academy of Sciencies, RANLP Summer School in Tzigov Chark), Republica Ceha(Karol University of Prague, Masaryk University of Brno), Danemarca (invited panelist at COLING, Copenhagen), Polonia (Warsaw University, Posnan University), Ungatia (Hungarian Academy of Sciences, tutorial at EACL conference in Budapest), Slovacia (International Laboratory-Bratislava), Elvetia (University of Geneve), Germania (Technical University of Berlin, University of Potsdam, University of Tubingen), Grecia (University of Patras), Italia (Institute for Computational Linguistics-Pisa, University of Florence, University of Venice), Luxemburg (invited by European Commission), Olanda (IPO - Utrecht), Portugalia (Calouste Gulbenkian Foundation in Lisbon), Republica Moldavaa (Academia de Stiinte Tehnicesi Academia de Stiinte Economice in Chisinau), Rusia (Tver NPO-TPS, Russian Academy of Sciences in Moscow), UK (Royal Society in London, Surrey University, University of Edinburgh, University of Birmingham, University of Mancester, University of Leeds, University of Sheffield), Franta (LIMSI-Paris, Xerox Researh Center - Grenoble, LORIA-Nancy, University of Strassbourg, Jussieu University of Paris (VII), UNESCO-Paris), SUA(George Mason University, Stanford Research Institute, University of California at Santa Cruz, University of Maryland, University of Pennsylvania, University of Texas, University of New Mexico).

Details:
Research Interests: Corpus Linguistics, Intelligent Computer Aided Language Learning, Machine Language Learning, Machine Translation, Natural Language Understanding, Natural Language Generation, Knowledge Representation.
Dr. Tufis is the main author of the first NL systems for Romanian SDLR, IURES, PC-IURES, PARADIGM and also the main author of TC-LISP.
Dr. Tufis and his colleagues developed the first Romanian standardized corpus (TEI/CES-conformant, annotated and validated up to the word level) and the associated language resources (the Romanian wordnet, DEX explanatory dictionary -CONCEDE conformant , a large wordform lexicon - MULTEXT conformant, Romanian paradigmatic morphology, tokenisation formal description, POS guessing rules for unknown words, tagsets for statistical disambiguation, etc. Dr. Tufis is the author of the tiered tagging approach to statistical POS disambiguation for large tagsets and of the reified alignment model for aligning of parallel texts;
Dr. Dan Tufiş has been invited as a visiting scientist and invited lecturer at different universities and research institutes from Europe, Asia, and North America (USA and Canada): Belgium (Vrije University of Brussels), Bulgaria (Bugarian Academy of Sciencies, RANLP Summer School in Tzigov Chark), Czech Republic (Karol University of Prague, Masaryk University of Brno), Denmark (invited panelist at COLING, Copenhagen), Poland (Warsaw University, Posnan University), Hungary (Hungarian Academy of Sciences, tutorial at EACL conference in Budapest), Slovakia (International Laboratory-Bratislava), Switzerland (University of Geneve), Germany (Technical University of Berlin, University of Potsdam, University of Tubingen), Greece (University of Patras), Italy (Institute for Computational Linguistics-Pisa, University of Florence, University of Venice), Luxembourg (invited by European Commission), The Nederlands (IPO - Utrecht), Portugal (Calouste Gulbenkian Foundation in Lisbon), Republic of Moldavia (Academy of Technical Sciences and Academy of Economic Sciences in Chisinau), Russia (Tver NPO-TPS, Russian Academy of Sciences in Moscow), UK (Royal Society in London, Surrey University, University of Edinburgh, University of Birmingham, University of Mancester, University of Leeds, University of Sheffield), France (LIMSI-Paris, Xerox Researh Center - Grenoble, LORIA-Nancy, University of Strassbourg, Jussieu University of Paris (VII), UNESCO-Paris), USA (George Mason University, Stanford Research Institute, University of California at Santa Cruz, University of Maryland, University of Pennsylvania, University of Texas, University of New Mexico). Several invited talks in Romania (at the Romanian Academy, most universities and many research institutes).

Publicații selectate:

* Dan Tufiş, Word Senses: The Stepping Stones in Semantic-Based Natural Language Processing., Springer-Verlag, Maglogiannis, I., Karpouzis, K., Bramer, M, Artificial Intelligence Applications and Innovations, 204, 2006.

* Dan Tufiş, Exploiting Aligned Parallel Corpora in Multilingual Studies and Applications., Springer-Verlag, Berlin Heidelberg, Ishida, T., Fussell, S.R., Vossen, P, Intercultural Collaboration I, LNCS 4568, 2007.

* Dan Tufiş, Svetla Koeva, Ontology-supported Text Classification based on Cross-lingual Word Sense Disambiguation, Springer-Verlag Berlin Heidelberg, F. Masulli, S. Mitra, and G. Pasi, WILF 2007, LNAI 4578, 2007.

* 1. Georgiana Puşcaşu, Adrian Iftene, Ionuţ Pistol, Diana Bãlaşa, Dan Tufiş, Alin Ceauşu, Dan Ştefãnescu, Constantin Orãşan, Iustin Dornescu, Radu Ion, Dan Cristea, A Cross-Lingual Romanian to English Question Answering System, Springer-Verlag, Peters, C.; Clough, P.; Gey, F.C.; Karlgren, J.; Magnini, B.; Oard, D.W.; de Rijke, M.; Stempfhuber, M., Lecture Notes in Computer Science, Springer-Verlag, 4730, 2007.