Lachlan McPheat

(he/him)
Picture of Lachlan in front of the St Pauls Cathedral in London

As of August 2024, I am a research associate at The Alan Turing Institute, working on the Robust Inference with Probabilistic Answer Set Programs Scaffolds for Large Language Models.

From May to August 2024 I was visiting the Bekki Lab at Ochanomizu University in Tokyo, studying Bekki's Dependent Type Semantics of natural language.

From June to September 2022 I interned at Quantinuum's Oxford office, working under Bob Coecke in graphical Discourse Representation Theory.

I finished my PhD at the PPLV , UCL Computer Science, supervised by Mehrnoosh Sadrzadeh. I studied methods of modelling referential phenomena in natural lanugage using distributional-compositional semantics. So far this involves defining categorical interpretations of variants of Lambek Calculus with modalities which resolve coreference syntactically and experimenting with the semantics of this resolution using a variety of composition and language models.

Broader research interests include: Transparency and Interpretability in AI/ML, Computational Linguistics, NLP, Distributional-Compositional Semantics, Categorical Semantics, Category Theory and Dependent Type Theory.

Research Activities

Teaching, Tutoring and Outreach

Factoids

Contact me by email at l (dot) mcpheat (at) UCL (dot) ac (dot) uk