The Department of Computational Linguistics is working on an automatic analysis of Polish text and computer dictionaries, that are basic tools to perform a text analysis.
Polish is a highly inflected language, therefore the basic tool must be an inflection dictionary (SFJP), i.e. a computer program, which can convert the inflection form, e.g. psa (gen. sing. ‘dog’) into head form pies (nom. sing. ‘dog’) and can assign the grammatical description (i.e. noun, animate, male) to the word.
The word identified by the inflection dictionary may bring more information, if described by semantic dependencies to other words. In the semantic dictionary (SSJP) we use both syntagmatic e.g. pies ‘dog’ ACTION szczekać ‘bark’ and paradigmatic dependencies, e.g. pies ‘dog’ superordinate – subordinate e.g. owczarek ‘shepherd dog’.
The algorithm that performs operations on a text content may also use the knowledge of natural word connections, e.g. that krzesło ‘chair’ is strongly associated with stół ‘table’, which implies that krzesło ‘chair’ may be treated as present in a semantic structure, even if lexically absent in a sentence, e.g. guests are sitting around the table. The knowledge of natural connections is stored in an experimentally built word association network (SSSJP).
The individual work is focused on computer science, linguistics, psycholinguistics, philosophy, poetics and rhetoric.
Since the year 2000 the Department is supervising an interdisciplinary studies on human-computer interaction called Elektroniczne Przetwarzanie Informacji (Electronic Information Processing, EPI).
Department hosted a conference NLPCS 2015 http://nlpcs2015.epi.uj.edu.pl.