I am Associate Professor in the Computational Linguistics Group at the University of Groningen, where I lead the InClow research group working on Interpretable, Cognitively inspired, Low-resource language models.

I am passionate about the statistical modeling of human languages, particularly in a multilingual context. My long-term goal is to design robust language processing algorithms that can adapt to the large variety of linguistic phenomena observed around the world. Among others, I work towards improving the quality of LLM generation tasks in `challenging' languages, and making state-of-the art NLP systems more interpretable. As a cross-disciplinary research enthusiast, I'm interested in enhancing research on human language processing or language evolution with computational modeling tools.
Last but not least, I enjoy observing, interacting with, and finding daily inspiration in my two daughters and their trilingual minds in the making.

My research is funded by various Dutch Research Council (NWO) grants. Since 2024 I lead a Vidi project aimed at improving low-resource language modeling, taking inspiration from child language acquisition insights. I am part of two national-consortium projects, both funded by NWA-ORC initiatives: Interpreting deep learning models for language, speech & music (InDeep) and Low Resource Chat-based Conversational Intelligence (LESSEN). I also supervise two China Scholarship Council (CSC)-funded PhD students working on the simulation of human patterns of language learning and change. My earlier research was funded by a Veni grant focusing on understanding and improving the encoding of linguistic structure in Neural Machine Translation models (2017-2021).

Want to know more about my work? Please visit our group website InClow.

News

  • [May 2025]   We have a brand new group website! Check it out at https://inclow-lm.github.io
  • [Mar 2025]   I gave a keynote talk at NoDaLiDa/Baltic-HLT 2025, titled "Not all Language Models need to be Large: Studying Language Evolution and Acquisition with Modern Neural Networks". Tallin was great and so was NoDaLiDa and its lively community of Nordic and Baltic NLP'ers!
  • [Nov 2024]   We have received an Outstanding Paper Award at the BabyLM Challenge (CoNLL'24 Shared Task) for our paper "BabyLM Challenge: Exploring the Effect of Variation Sets on Language Model Training Efficiency" with Akari Haga, Akiyo Fukatsu, Miyu Oba, and Yohei Oseki.
  • [Nov 2024]   Excited to see my Vidi project finally take off with dream team members Francesca Padovani (PhD) and Jaap Jumelet (postdoc).
  • [Oct 2024]   Paper accepted at CoNLL: "NeLLCom-X: A Comprehensive Neural-Agent Framework to Simulate Language Learning and Group Communication", with Yuchen Lian and Tessa Verhoef.
  • [Oct 2024]   Paper accepted at EMNLP: "Model Internals-based Answer Attribution for Trustworthy Retrieval-Augmented Generation", with Jirui Qi, Gabriele Sarti and Raquel Fernández [Check out the HuggingFace DEMO!]
    • Disclaimer for prospective students: I regularly receive emails from external students interested in my supervision. I do my best to reply each of them, but don't always manage. If I have paid research positions (PhD, postdoc) I'll always post them here, so no post => no position :-(
    As for research internships: I don't take students from external universities due to an already high supervision load.