I’m a PhD researcher at the University of Edinburgh, currently being supervised by Mark Steedman, Itamar Kastner, and András Bárány.
My research interests include the following:
Carslaw, I., Bárány, A., Kastner, I. & Steedman, M., (2026) “An LLM Investigation into Inherent and Structural Case Representation: a German Case Study”, Society for Computation in Linguistics 9(1). doi: https://doi.org/10.7275/scil.4161
Carslaw, I., Milton, S., Navarre, N., Qing, C. & Uegaki, W., (2025) “Automatic Extraction of Clausal Embedding Based on Large-Scale English Text Data”, Society for Computation in Linguistics 8(1): 12. doi: https://doi.org/10.7275/scil.3151
“LLMs hinder Linguistic Decolonisation” at LELPGC (the University of Edinburgh - June 2026)
“LLMs and Reflexive Binding” at North East Syntax Seminar (the University of Aberdeen - June, 2024)
“The Inherent-Structural Case Divide: An investigation into German morphological case representation” at LELPGC (the University of Edinburgh - June, 2025)
“LLMs and Reflexive Binding Intuitions: Do LLMs encode native speakers’ reflexive binding intuitions?” at the Form-meaning Mismatches in Spoken and Visual Communication Summer School (the University of Göttingen - August, 2024)
PhD - CDT in NLP at the University of Edinburgh (Current)
1st class MSc - Speech and Language Processing at the University of Edinburgh (2021-2022). Dissertation: “Low resource neural machine translation and generalised compositionality”
1st class BA (Hons) - Linguistics at UCL (2017-2021). Dissertation: “Parasitic Gaps: can they really be accounted for by sideward movement?”
I tutor on the following courses taught at the University of Edinburgh:
I also ran CLING (a computational linguistics reading group) from 2024-2025, and was on the organising committee for Language Lunch (an informal poster session for language related research by the postgrad community at Edinburgh) from 2023-2024.
Disclaimer: I do not give permission for this webpage to be data scraped nor do I give permission for any of the information on this page to be used in the process of training an LLM or ML model.