Prelucrarea limbajului natural | Natural language processing |
trul |
|||||
(Mathematics-Computer Science) |
|||||
(Computer Science) |
Cadre didactice indrumatoare | Teaching Staff in Charge |
Conf. Dr. TĂTAR Doina, dtatar@cs.ubbcluj.ro |
Obiective | Aims |
Prezentarea unuia dintre cele mai active domenii ale inteligentei artificiale, prelucrarea limbajului natural. In acest scop este prezentat modelul lingvistic de structura de atribute.
Deasemenea sunt prezentate problemele de semantica si de pragmatica care apar in acest domeniu. Acestea sunt importante atunci cand sunt abordate aspectele legate de dialogul (interfetele) om-calculator. |
Natural language processing is now accepted as one of the most studied and active field of Computer Science. The notion of feature structure as linguistic object stands on the base of
most recent approaches which are surveyed in this course. Also, the semantics and pragmatics of natural language processing which are central in human -computer interaction, text summarization, text generation and machine translation are presented. |
1. Structuri de atribute (FS) ca obiecte de reprezentare a cunostintelor lingvistice.
Ierarhii de tipuri.Relatia de subsumare pe multimea FS. Unificarea ca operatie pe multimea FS. Descriptori si legatura cu FS.Teoria deductiva a descriptorilor.Cel mai general FS care satisface un descriptor.Cazul variabilelor si legatura cu programarealogica. Notiunea de "potrivire" a tipurilor cu atributele.Notiunea de buna-tipizare, total -buna-tipizare. 2. Gramatici de unificare. Formalisme alternative: gramatici ale clauzelor definite,formalismul Patr, gramatici HPSG. Reguli de rescriere. Evaluari top-down si evaluari bottom-up. 3. Metode statistice de prelucrare a limbajului natural. Modelul Markov acoperit ( HMM). Drum de probabilitate maxima intr-un HMM, calculul probabilitatii unor secvente de intrare. Gramatici probabiliste. 4. Dezambiguare semantica. Dezambiguare supervizata. Dezambiguare bazata pe dictionare si "thesauri". Algoritme de dezambiguare client/server.Dezambiguare "bootstrepping". Dezambiguare nesupervizata. Aplicatii la categorizarea de text, sumarizarea de text si traducere automata. |
1. J.Allen : "Natural language understanding", Benjamin/Cummings Publ., 2nd ed., 1995.
2. B.Carpenter: "The logic of typed feature structures", Cambridge University Press,1992. 3. B.Carpenter: "ALE:The attribute logic engine.User's guide". Carnegie Mellon University,1994. 4. M.Johnson:"Attribute-value logic and the theory of grammar",CSLI,1988. 5. D.Jurafsky, J.Martin: "Speech and language processing", Prentice Hall, 2000. 6. A.Lecomte: "Grammaire et theorie de la preuve. Une introduction" , T. A. L. ,1996, vol 37, no.2, pp 3-37. 7. C.Manning, H.Schutze: "Foundation of statistical natural language processing", MIT, 1999. 8. G.Morrill:"Type LogicalGrammar.Categorial Logic of Signs", Kluwer Academic Publishers, 1994. 9. S.J.Russell, P.Norvig: "Artificial intelligence.A modern approach", Prentice-Hall International, 1995. 10. M.Shieber: "Introduction to unification-based approaches to grammars", CSLI,1986. 11. D.Tatar: "Inteligenta artificiala: demonstrare automata de teoreme, prelucrarea limbajului natural", Editura Microinformatica, 2001. 12. D.Tatar: "Unification Grammars in Natural Language Processing", in "Recent topics in mathematical and computational linguistic", ed. C. Martin-Vide, G. Paun, Editura Academiei, 2000, pg. 289-300 |
Evaluare | Assessment |
Examenul este oral, cu subiecte din intreaga materie. |
The examination is by oral exam, with the subjects from all the matter. |