
Academic lab · Riga, Latvia
AiLabUniversity of Latvia IMCS
The AI Laboratory at the University of Latvia builds open Latvian language resources, from the Tezaurs lexicon to a multilayer corpus and Latvian BERT models.
Riga, Latvia
- Founded
- 1992
- Type
- Academic lab
- Based in
- Riga, Latvia
- Featured projects
- 4
About
The Artificial Intelligence Laboratory belongs to the Institute of Mathematics and Computer Science at the University of Latvia in Riga and has conducted research since 1992. Its work covers natural language processing and machine learning, including Latvian speech recognition and synthesis, syntactic and semantic analysis and language models.
The lab maintains a set of public Latvian resources, among them Tezaurs.lv, Korpuss.lv, the LV-PIPE tools, CLARIN-LV and Balsutalka.lv. Its FullStack-LV repository offers a multilayer corpus with Universal Dependencies, FrameNet, PropBank and AMR layers, released under a Creative Commons licence.
Funding comes from the European Union, the European Regional Development Fund and Latvian ministries and councils. The lab has partnered with the news agency LETA, Deutsche Welle and the National Library of Latvia.
Mission and ambitionsThe lab states that it develops resources and technologies that help Latvian thrive in a multilingual environment. It publishes its tools and data openly for researchers and developers.
Known for
- 01
Tezaurs Latvian lexical database
- 02
FullStack-LV multilayer corpus
- 03
LVBERT Latvian language model
- 04
SemEval 2016 and 2017 results
Milestones
- 2015
Tezaurs repository
The open Latvian lexical database repository was created on GitHub in April 2015.
- 2016
SemEval 2016 semantic analysis
The lab reports the best result in the semantic analysis task at SemEval 2016.
- 2017
SemEval 2017 text generation
The lab reports the best result in the text generation task at SemEval 2017.
- 2018
FullStack-LV
A multilayer Latvian corpus with UD, FrameNet, PropBank and AMR layers was started and published at LREC 2018.
- 2020
LVBERT
A pretrained Latvian BERT model repository was created in July 2020.
- 2023
MultiCoNER 2023
The lab placed third among European teams in the named entity recognition task.
Open source
The largest open lexical database of Latvian (65 stars)
Full Stack of Latvian Language Resources for Natural Language Understanding (NLU) and Generation (NLG) (16 stars)
Pretrained contextualized transformer-based state-of-the-art language model for Latvian (14 stars)
Studiju kursu "Valodu tehnoloģiju pamati" un "Valodu tehnoloģiju lietojumi" praktisko darbu materiāli (10 stars)
Notable releases
Openness
Language resources and models are public on GitHub.
People
- Normunds GrūzītisCoordinator of the FullStack-LV project
- Pēteris PaikensResearcher and member of the LUMII-AILab GitHub organization
- Lauma PretkalniņaResearcher, enhanced UD from the hybrid treebank
Listed from the lab's own pages and public profiles.
Contact and community
Contact via the address printed on the lab website.
Sources
Compiled from the lab's public pages, Hugging Face and GitHub. Last checked 2026-11-17. Something wrong or out of date? Tell us on Discord.



