Instituto de Linguística Teórica e Computacional

Instituto de Linguística Teórica e Computacional
Name	Instituto de Linguística Teórica e Computacional
Native name	Instituto de Linguística Teórica e Computacional
Established	1990
Type	Research institute
City	Lisbon
Country	Portugal

Instituto de Linguística Teórica e Computacional is a Portuguese research institute based in Lisbon focusing on theoretical linguistics and computational linguistics. The institute conducts research spanning phonology, syntax, semantics, corpus linguistics and natural language processing while engaging with universities, research councils and international consortia. It participates in national and European funding frameworks and contributes to standards, corpora and open-source tools used by academic and industry partners.

History

The institute was founded in the early 1990s during reforms associated with Universidade de Lisboa, Universidade Nova de Lisboa, Fundação para a Ciência e a Tecnologia, European Union research initiatives and the expansion of computational resources in Portugal. Early collaborations involved researchers from Universidade de Coimbra, Universidade do Porto, Instituto Superior Técnico, Faculdade de Ciências da Universidade de Lisboa, and international visitors from University of Cambridge, Massachusetts Institute of Technology, Stanford University, University of Pennsylvania and Université Paris Diderot. Throughout the 2000s the institute engaged with projects funded by the European Commission, Horizon 2020, European Research Council and bilateral agreements with institutions such as University of Oxford, Max Planck Institute for Psycholinguistics, Université Paris-Saclay and University of Edinburgh. Its development paralleled growth at laboratories like INRIA, MIT Computer Science and Artificial Intelligence Laboratory, German Research Center for Artificial Intelligence, and collaborations with museums and cultural bodies including Museu Nacional de Arte Antiga. The institute's timeline includes participation in landmark initiatives connected to Text Encoding Initiative, ISO standards and multilingual resources linked to Common Language Resources and Technology Infrastructure consortia.

Research Areas and Projects

Research spans theoretical and computational domains, connecting work on Noam Chomsky-inspired syntax frameworks, Generative Grammar-related debates, interfaces with Montague Grammar, corpus-driven studies with links to Brown Corpus, British National Corpus, Corpora of Contemporary Portuguese and computational models influenced by methods from Hidden Markov Model, Conditional Random Field and Transformer (machine learning model). Projects address speech technologies related to Festival (software), HTK (Hidden Markov Model Toolkit), and text technologies interoperable with TEI-annotated corpora. The institute has contributed to machine translation programs influenced by research from Google Translate, statistical paradigms from IBM Watson era research, and neural methods following architectures from OpenAI, DeepMind, and Facebook AI Research. Work on lexical databases connects with resources like WordNet, Wiktionary, Universal Dependencies and named-entity resources used in shared tasks organized by Conference on Computational Natural Language Learning, Association for Computational Linguistics and European Language Resources Association. Projects include speech recognition, syntactic parsing, semantic role labeling, sentiment analysis, and low-resource language documentation linked to fieldwork traditions of Linguistic Society of America and archives such as The Endangered Languages Archive.

Organization and Structure

The institute is organized into thematic groups aligned with units found in research organizations such as Instituto de Engenharia de Sistemas e Computadores, Centro de Linguística da Universidade Nova de Lisboa, Max Planck Society institutes and sections similar to those at National Institutes of Health research centers. Leadership roles have corresponded with faculty holding appointments at Universidade de Lisboa, Universidade do Porto, Universidade de Coimbra and guest positions from scholars affiliated with University of California, Berkeley, Princeton University, Yale University and University of Chicago. Administrative and technical teams coordinate grant management with bodies like Fundação Calouste Gulbenkian and national agencies such as Ministério da Ciência, Tecnologia e Ensino Superior (Portugal). The institute maintains computational infrastructure paralleling setups at ELIXIR, CLARIN, GRID and cloud collaborations with providers used by European Open Science Cloud participants.

Publications and Resources

The institute publishes articles in venues such as Journal of Linguistics, Computational Linguistics (journal), Transactions of the Association for Computational Linguistics, Language, Lingua, and conference proceedings for ACL (conference), EMNLP, COLING, EACL and LREC. It produces corpora, lexica and annotation guidelines interoperable with ISO 639-3 identifiers and contributes datasets to repositories used by Zenodo and archives in partnership with Biblioteca Nacional de Portugal. Software outputs have included tools distributed under licenses common to projects from The Apache Software Foundation and GNU Project ecosystems and integrations with platforms like GitHub and GitLab. The institute also issues technical reports, working papers and monographs comparable to collections from Cambridge University Press and Oxford University Press authors.

Collaborations and Partnerships

Collaborative partners include universities and research centers such as University of Lisbon, University of Porto, University of Coimbra, Universidade Nova de Lisboa, Max Planck Institute for Psycholinguistics, Inria, University of Cambridge, Massachusetts Institute of Technology, Stanford University, University of Edinburgh and industry partners in technology sectors comparable to Microsoft Research, Google Research, Amazon Web Services, IBM Research and Facebook AI Research. The institute engages in European projects with partners including European Commission, Horizon Europe networks, CLARIN, ELRA and national cultural institutions like Arquivo Nacional Torre do Tombo. Long-term bilateral links have been maintained with archives and research centers in Brazil, Mozambique, Angola and Lusophone academic institutions such as Universidade Federal do Rio de Janeiro and Universidade de São Paulo.

Teaching and Training Programs

The institute contributes to postgraduate programs at Universidade de Lisboa, Universidade Nova de Lisboa, Universidade do Porto and supervises doctoral candidates enrolled at institutions comparable to MIT, Stanford University and University of Cambridge. Training includes summer schools modeled on programs by European Summer School in Logic, Language and Information, workshops linked to Association for Computational Linguistics events, and professional short courses for partners in industry similar to training offered by Coursera and edX collaborations. It also hosts internships and visiting researcher fellowships funded through schemes like Marie Skłodowska-Curie Actions and bilateral exchange programs with centers such as Max Planck Society and CNRS.

Category:Linguistics research institutes