Projects
Recent projects
In Natural Language Processing Centre (NLPC), I am working on the following projects:
- Language Services: API Collection of NLPC services: NLP Services
- TextPerience: experience the text, demo web application: http://nlp.fi.muni.cz/projects/textperience/
- foreign (mainly English) inclusions in Czech text and their detection in corpora
- textual entailment and paraphrasing generation for Czech
- see my PhD thesis
- free dataset of triples text-hypothesis-annotation
- named entity recognition for Czech
- http://nlp.fi.muni.cz/projects/ner/
- made initially by me, then enhanced by Lubomír Sedlář (his thesis, in Czech)
- topic recognition for Czech
- http://nlp.fi.muni.cz/projects/declension/
Former projects
- predictive writing
- http://nlp.fi.muni.cz/projects/predictive/predictive-ui/
- predictive writing based on large models (150,000,000 n-grams)
- the general model can be enhanced by user-defined model
- intelligent phrase search & replace
- http://nlp.fi.muni.cz/projects/phrase_replace
- input a text, insert a search noun phrase in nominative and a replace noun phrase in nominative
- so far it is experimental but the aim is to build an application that can help consistency in large technical texts
- other modules concerning the project Watsonson:
- http://nlp.fi.muni.cz/projects/watsonson/modules/index.py
- sentence generation in Czech
- pronominal anaphora resolution for Czech
- corpus-based chunk scoring: it is believed that chunks that have higher score (and thus appear more often in corpora) sound more natural to humans
- X-plain game and X-plain associative network (game with a purpose)
- similar to Verbosity game but one player is human, the other is the program
- the result is an associative network
- automatic declension of noun phrases, prepositional phrases, and adjective phrases
- visualization of networks: project Visual Browser: http://nlp.fi.muni.cz/projects/visualbrowser
- EuDML: European Digital Math Library, see https://eudml.org/
Backlinks: English