Harvey Donnelly

Website: https://harveyrdonnelly.github.io/

Research keywords: Computational neuroscience, large vision-language models, deep reinforcement learning, cognitive modelling

Bio:

Harvey is from the city of Belfast in the north of Ireland. He was awarded the Lester B. Pearson International scholarship in 2021, a full-ride merit-based scholarship towards the completion of an undergraduate degree at the University of Toronto. He graduated in 2025 with a BSc. (Hons.) Specialist in Computer Science, focus in Artificial Intelligence, and a Major in Cognitive Science. Harvey’s undergraduate studies were guided by an interdisciplinary perspective, through which he developed a foundational background in machine learning, natural language processing and computational neuroscience. During his final year of studies, he worked as a research assistant at the Cognitive Neuroscience and Sensorimotor Integration (CoNSens) Lab at the University of Toronto, where he developed deep reinforcement learning networks to model task feature integration for sensorimotor tasks. He also conducted research at the Computational Linguistics and Natural Language Processing (CL&NLP) group, investigating novel methods for modelling automatic speech recognition error to enhance downstream spoken language understanding.

PhD research:

Harvey’s research interests concentrate on developing machine learning techniques and computational models to explain cognition in both humans and machines. He is particularly interested in understanding shared representations between vision and language, how these are shaped by optimising for multiple task objectives and how emergent representations can be comparatively analysed between the brain and large vision-language models (LVLMs). He aims to develop frameworks for naturalistic optimisation and benchmarking of cognitive capacities, leveraging methods in deep reinforcement learning and natural language processing.

Supervisors: Benjamin Peters, Frank Keller