Universal natural language understanding
The project develops methods for constructing machine-readable semantic representations of natural language.
About the project
Computers have a hard time understanding linguistic meaning. If you search for "Bordeaux wines with merlot" and "Bordeaux wines without merlot" you will get the same search results. This is because "with" and "without" are very common words that can be found on any web page.
In this project, we try to help the computer understand what you mean and therefore realize that when you look for "Bordeaux wines without merlot" you want information that can be found on web pages where "merlot" does not occur.
We will develop a method for constructing machine-readable semantic representations of natural language, based on syntactically parsed texts.
- Develop an overall, broadly language-independent framework for converting syntactic representations in the Universal Dependencies format to semantic representations
- Adapt this framework to applications for Norwegian
- Develop and implement a theory of the syntax-semantics interface for dependency grammars
- Learn tree transducers for normalization of parser output from gold standard semantic annotations
Findlay, Jamie Yates & Haug, Dag Trygve Truslew (2021). How useful are Enhanced Universal Dependencies for semantic interpretation? In Osenova, Petya (Eds.), Proceedings of the Sixth International Conference on Dependency Linguistics (Depling, SyntaxFest 2021). Association for Computational Linguistics. ISSN 9781955917148. p. 22–34. Full text in Research Archive
Findlay, Jamie Yates (2021). Meaning in LFG. In Arka, I Wayan; Asudeh, Ash & King, Tracy Holloway (Ed.), Modular Design of Grammar: Linguistics on the Edge. Oxford University Press. ISSN 9780192844842. p. 340–374.