Compositional vector-based semantics for Dutch

Project

Description

`A composition calculus for vector-based semantic modelling with a localization for Dutch’ is a research project funded by the Dutch Research Council (NWO 360-89-070, 2017-2022). Focusing on empirical data from Dutch, the project aims to combine a vector-based modelling of lexical semantics with a typelogical compositional account of how word meanings are combined to obtain interpretations for larger linguistic units. The project investigates this approach to compositionality with the objective of providing a collection of computational tools and resources for the compositional distributional study of Dutch.

People

Project team:

– Michael Moortgat (PI)
– Konstantinos Kogkalidis
– Gijs Wijnholds
– Giuseppe Greco (2017-2020)

External collaborators:

– Mehrnoosh Sadrzadeh (UCL)
– Richard Moot (LIRMM Montpellier)

Internships:

– Giorgos Tziafas (RU Groningen)
– Valentin Richard (ENS Paris-Saclay)
– Marieke Hofslot (UU)

Tools/Resources

Publications

  • Kogkalidis, Konstantinos and Wijnholds, Gijs. 2022. “Discontinuous Constituency and BERT: A Case Study of Dutch.” In Findings of the Association for Computational Linguistics 2022, 10.18653/v1/2022.findings-acl.298
  • Wijnholds, Gijs, and Michael Moortgat. 2021. “SICK-NL: A Dataset for Dutch Natural Language Inference.” In Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics (EACL), edited by Paola Merlo, Jörg Tiedemann, and Reut Tsarfaty, 1474–79. Association for Computational Linguistics. https://doi.org/10.18653/v1/2021.eacl-main.126.
  • Kogkalidis, Konstantinos, Michael Moortgat, and Richard Moot. 2020a. “ÆTHEL: Automatically Extracted Typelogical Derivations for Dutch.” In Proceedings of the 12th Language Resources and Evaluation Conference, LREC 2020, Marseille, 5257–66. European Language Resources Association. https://www.aclweb.org/anthology/2020.lrec-1.647/.
  • ———. 2020b. “Neural Proof Nets.” In CoNLL2020, Proceedings of the 24th Conference on Computational Natural Language Learning. Association for Computational Linguistics. https://arxiv.org/abs/2009.12702.
  • Moortgat, Michael, Mehrnoosh Sadrzadeh, and Gijs Wijnholds. 2020. “A Frobenius Algebraic Analysis for Parasitic Gaps.” Journal of Applied Logics 7 (5): 823–52.
  • Tziafas, Giorgos, Konstantinos Kogkalidis, Gijs Wijnholds, and Michael Moortgat. 2021. “Improving BERT Pretraining with Syntactic Supervision.” CoRR abs/2104.10516. https://arxiv.org/abs/2104.10516.
  • Wijnholds, G., M. Sadrzadeh, and S. Clark. 2020. “Representation Learning for Type-Driven Composition.” In CoNLL2020, Proceedings of the 24th Conference on Computational Natural Language Learning. Association for Computational Linguistics.
  • Kogkalidis, Konstantinos, Michael Moortgat, and Tejaswini Deoskar. 2019. “Constructive Type-Logical Supertagging with Self-Attention Networks.” In Proceedings of the 4th Workshop on Representation Learning for NLP, RepL4NLP@ACL 2019, Florence, 113–23. Association for Computational Linguistics. https://doi.org/10.18653/v1/w19-4314.
  • Kogkalidis, Konstantinos. 2019. “Extracting and Learning a Dependency-Enhanced Type Lexicon for Dutch.” Master’s thesis, Utrecht University. https://arxiv.org/abs/1909.02955.
  • Greco, Giuseppe, Fei Liang, Michael Moortgat, and Alessandra Palmigiano. 2020. “Vector Spaces as Kripke Frames.” Journal of Applied Logics 7 (5): 853–73.
  • Moortgat, Michael, and Gijs Wijnholds. 2017. “Lexical and Derivational Meaning in Vector-Based Models of Relativisation.” In Proceedings of the 21st Amsterdam Colloquium, edited by Alexandre Cremers, Thom van Gessel, and Floris Roelofsen, 55–64. Universiteit van Amsterdam.