Henry Gouk

Henry Gouk

I'm a Lecturer (Assistant Professor) in Machine Learning at the School of Informatics within the University of Edinburgh, where I co-lead the Bayesian and Neural Systems group. I also hold a Research Fellowship from the Royal Academy of Engineering. Prior to this, I was a postdoctoral researcher working with Prof Timothy Hospedales at the University of Edinburgh, and before that I completed my PhD at the University of Waikato under Prof Bernhard Pfahringer and Prof Eibe Frank.

My research interests can be broadly categorised as Artificial Intelligence Engineering: the design and analysis of methods for economically building reliable AI systems. I am particularly interested in thinking about machine learning models as components in a broader system, and much of my recent work has centred on the following challenges:

Robustness to distribution shift: how can we make machine learning systems that continue to work well when the world changes?
Data-efficient learning: how can we make machine learning systems from small amounts of data?
Computational efficiency: how can we make machine learning systems that run quickly and cheaply?

I am always looking for talented students and researchers to join my group. Please get in touch if you are interested in working with me!

Team

Philipp Bomatter

Philipp Bomatter

PhD Student

Philipp's research focuses on developing machine learning models for analysing EEG data, with a particular emphasis on generalising across subjects and recording sessions.

Jack Geary

Jack Geary

PhD Student

Jack is investigating how game-theoretic formulations of distribution shift can be integrated with machine learning algorithms to improve robustness and reduce the need for retraining.

Cameron Barker

Cameron Barker

PhD Student (with Elliot Crowley)

Cameron is investigating how modern foundation model training and inference can be made more efficient through hardware-aware algorithmic innovations.

Taehoon Kim

Taehoon Kim

PhD Student (with Timothy Hospedales)

Taehoon is developing methods for robust adaptation of modern machine learning models to distribution shifts, with a focus on applications in computer vision and natural language processing.

Sanghyeok Choi

Sanghyeok Choi

PhD Student (with Esmeralda Whitammer)

Sanghyeok is investigating how large language models can be used as a source of prior knowledge for causal discovery and inference problems.

Selected Publications

Model merging is secretly certifiable: Non-vacuous generalisation bounds for low-shot learning

Model merging is secretly certifiable: Non-vacuous generalisation bounds for low-shot learning

Taehoon Kim, Henry Gouk, Timothy Hospedales

UAI, 2026

Weight-space learning for certifiable few-shot transfer learning

Weight-space learning for certifiable few-shot transfer learning

Fady Rezk, Royson Lee, Henry Gouk, Timothy Hospedales, Minyoung Kim

ICML, 2026

Test-Time Alignment of Text-to-Image Diffusion Models via Null-Text Embedding Optimisation

Test-Time Alignment of Text-to-Image Diffusion Models via Null-Text Embedding Optimisation

Taehoon Kim, Henry Gouk, Timothy Hospedales

CVPR, 2026

Is Limited Participant Diversity Impeding EEG-based Machine Learning?

Is Limited Participant Diversity Impeding EEG-based Machine Learning?

Philipp Bomatter, Henry Gouk

NeurIPS, 2025