Research

The group conducts research on core problems in natural language processing:


how language conveys information,
how information can be extracted from text,
how latent structure can be learned from observed data.

Our work combines empirical and theoretical approaches. We work on projects in various application domains with partners in the academia, industry and government organizations.
PULS

Analysis of big-data streams of news media. Information Extraction: finding facts and events in text, and reasoning over extracted data. 

  • Methods: neural networks, supervised and weakly-supervised machine learning. 
  • Domains: general news, business intelligence, epidemiological surveillance, cross-border security and crime.
  • Collaboration: Please see project page for partners.
  • Funding: Tekes/BusinessFinland, European Commission
Revita

Computational modeling to support language learning.

Revitalization of endangered languages from the Finno-Ugric, Turkic, and other language families.

  • Collaboration: University of Helsinki Department of Modern Languages, Department of Finnish, Finno-Ugrian and Scandinavian Studies, YLE, Opetushallitus, University of Jyväskylä, Università degli Studi di Milano
  • Funding: Academy of Finland, Project FinUgRevita.
SIGSLAV

SIGSLAV, the Special Interest Group on Slavic Natural Language Processing, is dedicated to computational linguistics for Slavic languages. The purpose of SIGSLAV is to promote interest in basic and applied research on natural language processing (NLP) in all Slavic languages, including in areas such as: morphological analysis and generation, syntactic and semantic tagging, named-entity recognition, information extraction, co-reference resolution, question-answering, information retrieval, text summarization and machine translation.