We study how computers understand and produce human language, and how to make language technology reliable and useful in practice.
We are the natural language processing (NLP) group of IDSIA, the Dalle Molle Institute for Artificial Intelligence (USI-SUPSI) in Lugano, Switzerland. We combine basic research, mostly funded by the Swiss National Science Foundation, with applied projects together with hospitals and companies.
Research themes
All research →Health and clinical text
Language technology for hospitals and health research: clinical notes, drug safety and patient privacy.
Languages, dialects and multilinguality
How language models deal with dialects, scripts and less-resourced languages, and how to make them fairer across languages.
Understanding and trusting language models
When language models get things wrong, and how to tell: hallucinations, machine-generated text, uncertainty.
Knowledge from scientific literature
Turning scientific papers into structured, searchable knowledge, and making science easier to read.
News
All news →- New group leader Oscar William Lithgow-Serrano is the new leader of the NLP group. We thank Fabio Rinaldi, who founded the group in 2019 and led it until 2026.
- New SNF project TaxoMine starts in March Our new project on mining taxonomic knowledge from the scientific literature starts in March 2026 (working title: TaxoMine). We have two open positions.
- Award at EMNLP 2025 Our paper on tokenization and representation biases in multilingual models for dialects received a Senior Area Chair award at EMNLP 2025.
- New SNF project approved We will continue our collaboration with Robert Waterhouse on extracting taxonomic information from the scientific literature.
Current projects
All projects →- BM-TOK How to build the vocabulary of a multilingual language model so that less-resourced languages and non-Latin scripts are represented fairly.
- AutoDischarge Helps doctors write discharge letters faster by drafting them from the patient record, for the doctor to check and complete.
- UBS–IDSIA AI Lab The NLP group participates in the collaboration across different NLP-related projects.
- GURU (ENERGENIUS) A conversational interface where citizens ask in natural language about energy savings, costs and regulation.
Selected publications
All publications →- Script Choice in LLMs: Evidence for Late-Layer Commitment
- CANE is a Format: Rethinking Output Formats for Few-Shot LLM-Based Named Entity Recognition
- Are Language Models Script-Aware?