Hello! We're the NLP and Society lab, part of the Natural Language Processing group at the Department of Information and Computing Sciences at Utrecht University (NL). We're interested in:
- Analyzing and modeling the relation between language and society using computational methods (computational sociolinguistics).
- The development of NLP methods to study social phenomena (NLP and computational social science). We focus on the analysis of online conversations and measurement in the social sciences.
- Techniques for more robust and explainable NLP.
People
Melody Sepahpour-Fard
PhD student
Based at the University of Limerick
(co-advised with Michael Quayle)
Selected recent research
Measuring the instability of fine-tuning (ACL 2023), using the construct validity framework to evaluate the quality of text embeddings (EPJ Data Science 2022), learning style representations (Workshop on Representation Learning for NLP, 2022), algorithmic fairness (AIES 2023).
Pre-print: FTFT: efficient and robust Fine-Tuning by transFerring Training dynamics (comments welcome!)
Software
See our GitHub repository!
Funded projects
NWO Veni (2020-2024), NWO digital society 'Digitale samenleving - de geïnformeerde burger' (2020-2024), UU Research IT innovation fund 'PROVEE: Progressive Explainable Embeddings'