Attestation committee
Accreditation committee
Expert committee
Dispositions, instructions
Normative acts
MANPOWER TRAINING
Nomenclature
Institutions
Scientific councils
Seminars
Theses		 Scientific advisers
Scientists
Doctoral students
Postdoctoral students
CNAA logo

 română | русский | english


Extensible Techniques of Syntactic – Semantic Interpretation of Texts in Natural Language


Author: Creţu Sergiu
Degree:doctor of physics and mathematics
Speciality: 01.05.04 - Mathematical modelling, mathematical methods, software
Year:2007
Scientific adviser: Dumitru Todoroi
doctor habilitat, professor
Scientific consultant: Anatol Popescu
doctor habilitat, professor, Technical University of Moldova
Institution:
Scientific council:

Status

The thesis was presented on the 23 November, 2007
Approved by NCAA on the 20 December, 2007

Abstract

Adobe PDF document – 0.29 Mb / in romanian

Thesis

CZU 004.822(043)

Adobe PDF document 0.90 Mb / in romanian
134 pages

Keywords

Knowledge base, database, linguistic competence, corpus, gloss, artificial intelligence, interpretation, syntactic – semantic interpretation, information understanding, formal language, natural language, syntax, semantics, denotation semantics, semantic networks, extensible techniques.

Summary

This study represents a research concerning the principles and methods of elaboration of syntactic-semantic interpretation system (SSI) of texts in natural language (NL). When we say syntactic-semantic interpretation of texts in a NL, we mean the process of analyzing and extracting information contained in NL, realized according to a precise semantic context, assured by an adequate knowledge base.

The result of syntactic-semantic interpretation represents a system of glosses – adapted fragments of text in NL, which forms the semantic and pragmatic explication of the interpreted text.

Our research results from the necessity of processing large amounts of information given in the NL.

A developed ISS-GLOS system assures the efficient identification of information from the processed texts (manually, in LBC or in an automatic way) and its interpretation under the format of glosses. The process of accessing, modification, elimination of the stored information is effectuated using the same system. The efficiency of the given operations is in relation with the semantic principles of saving information.

The importance of this research consists in the elaboration of a new theoretical and practical background, which permits the organization of efficient systems of saving information with complex semantics and its future utilization in the syntactic-semantic interpretation of texts in a NL.

We obtained the following results:

  1. The axiomatic model of representation of information contained in the NL was elaborated. This result was obtained interpreting the grammatical categories. Using this model we formulated some requests towards the needed knowledge base. The process of synthesizing the fragment of the analyzed text is based upon morphological information, syntactic information and semantic information of the given text.
  2. Using the semantic, syntactic and pragmatic competences of the speaker, the notion of syntactic-semantic interpretation of texts in the NL was formulated and the structure of the knowledge base was précised. In order to administrate the knowledge base, a new formal language LBC was elaborated. The conceptual basis for this language is the axiomatic model of representing information contained in the NL.
  3. An interpretative model of giving sense to the texts in NL was elaborated. The translated and transition semantic network as a tool of syntactic-semantic interpretation of texts in the NL was defined. In order to apply an algorithm for the SSI of texts in the NL, the notion of semantic translator with stack (theorem 2.1.1) was introduced. A non-deterministic model of interpretation was proposed. The translated, transition networks were completed with the notion of weight.
  4. We have formulated two theses, which stipulate that the obtained result can be described by a multi-sort algebra and the process of giving sense to the syntactical constructions can be reduced to the homomorphism defined by the algebraic derivers. The scheme of representing the result was elaborated, using denotative techniques.
  5. A set of operations and was elaborated and the concept of virtual vocabulary – as method of organizing and extending the knowledge base has been formulated.
  6. The final result was considered to be a gloss.