Finite automata and natural-language lexicons

Franco-Brazilian Project CAPES/COFECUB 279/99

This project is dedicated to several aspects of the construction, storage and exploitation of natural-language lexicons. Some of these aspects are dependent and others independent of the lexicon's language. Our common background comprises several techniques. The objective of the project is to exploit more recent methods and ideas that have been tested only on one of the two languages, French and Portuguese, or that can be adapted in order to solve other problems.
Expected results include a lexicon-based lexical tagger of texts in Brazilian Portuguese, and progress in:

spelling correction,

lexicon compression and

construction, extension and maintenance of electronic lexicons.

Beginning: 1999
Duration: 2 years