Natural Language Processing


Natural Language Processing (NLP) consists in the development of models and algorithms for the simulation of the human linguistic process.
Major applications of this area have the following goals:
(i) to permit the man-machine communication, e.g. speech recognition/understanding, interfaces in natural language and so on;
(ii) to improve the man-man communication, e.g. automatic message classification, information extraction, generation of directives and machine translation.

Among the methodologies used in this research field are the theory of algorithms and formal languages, information theory, probability calculus and formal linguistics.

Major interest of the research group is the individuation of techniques for the automatica development of systems for NLP which have a degree of accuracy of the same level as systems produced by humans. Such techniques are usually based on the automatic analysis of set of big dimension texts.
Among the topics of the group resarach are at present:

  • Part-of-Speech (POS) tagging:
  • this application consists in assigning the syntactic cathegory to every word of a text, solvin the possible ambiguities on the basis of the context.

  • Parsing:
  • automatic syntactic analysis of the sentence; this phase is preliminary for each application based on the automatic understanding.

  • Information Extraction:
  • given a text on a specific domain, information following a scheme will be synthesized.

  • Machine Translation (MT):
  • automatic translation of special types of texts, as technical instructions, announcements, web pages and so on.
  • Mathematical Wordnet:
  • : dictionary of mathematical terms with hiponymous and hiperonymous relations.