Aarne Talman

About

I am a Visiting Scholar in language technology at the University of Helsinki, where I also teach the course Large Language Models and Generative AI for NLP. Alongside research, I work as Senior Manager, Software Development at AMD Silo AI.

I received my PhD in language technology from the University of Helsinki in 2024. Before that I studied computational linguistics and formal grammar at King’s College London and philosophy at the London School of Economics.

Research

Natural language understanding

Whether models that score well on language understanding benchmarks capture meaning, or exploit shortcuts in the data. My doctoral work showed that models can keep much of their performance even when benchmark data is deliberately corrupted, and that results often fail to carry over from one benchmark to another.

Evaluating generative language models

I am one of the organisers of ELOQUENT, a lab of shared tasks at CLEF for assessing the quality of generative language models, and have been involved since its first edition in 2024.

Formal grammar and semantics

Grammar formalisms, formal semantics and what they can tell us about the abilities and limits of neural language models.

Publications

Teaching

• Large Language Models and Generative AI for NLP, Master’s course, University of Helsinki, every autumn since 2024, with Jussi Karlgren and Dmitry Kan.
• Natural Language Understanding and Representation Learning, Master’s course, University of Helsinki, 2020, with Alessandro Raganato.
• Supervision of Master’s theses in language technology.

Contact

Email

aarne.talman@helsinki.fi

CV

Curriculum vitae (PDF)

ORCID

0000-0002-3573-5993

GitHub

aarnetalman

Research portal

University of Helsinki