Bristol School of Mathematics to make strong showing at world-leading conference on Artificial Intelligence

1 October 2025

The work of several statisticians and mathematicians from the University of Bristol has been recognised in the announcement of accepted papers for the Thirty-Ninth Annual Conference on Neural Information Processing Systems (NeurIPS 2025).

Amongst those scheduled to present papers at the prestigious NeurIPS 2025 event - which will be held at venues in both San Diego and Mexico City from 30 November to 07 December - three are Compass CDT students, one is a recent graduate of our BSc in Data Science with a Year in Industry, and one is a postdoctoral researcher on the INFORMED-AI hub.

All those presenting are members of the Institute for Statistical Sciences, or are closely associated with it, such as alumni, and have worked with its staff and students. Some papers were also co-authored or supported by academics across Mathematics, Computer Science, and Engineering Mathematics - testament to the University’s collaborative and cross-disciplinary approach to research.

Professor Nick Whiteley, Head of the Institute for Statistical Sciences and Director of the Compass CDT, both part of the School of Mathematics, said:

“Publishing research in the proceedings of the NeurIPS conference is extremely competitive. Six acceptances this year illustrates the strength of the Institute for Statistical Science in AI and machine learning, and I'm delighted to see Compass students playing a central role in authorship of these papers.”

The papers that have been accepted for the conference proceedings (Bristol staff and student profiles hyperlinked) include:

‘A Principle of Pre-Strategy Intervention for Multi-Agent Reinforcement Learning’
- Anjie Liu · Jianhong Wang Samuel Kaski · Jun Wang · Mengyue Yang

‘Conditional Distribution Compression via the Kernel Conditional Mean Embedding’
- Dominic Broadbent Nick Whiteley · Robert Allison · Tom Lovett

‘Direct Fisher Score Estimation for Likelihood Maximization’
- Sherman Khoo · Yakun Wang · Song Liu · Mark Beaumont

‘Distributional Training Data Attribution’
- Bruno Mlodozeniec · Isaac Reid · Sam Power David Krueger · Murat Erdogdu · Richard Turner · Roger Grosse

‘Missing Data Imputation by Reducing Mutual Information with Rectified Flows’
- Jiahao Yu · Qizhen Ying · Leyang Wang · Ziyue Jiang · Song Liu

‘Scale-invariant attention’
- Ben Anson Xi Wang · Laurence Aitchison

Professor Oliver Johnson, Head of the School of Mathematics and Professor of Information Theory, said:

“Mathematics underlies all artificial intelligence technologies and is key to its future development. We are so pleased to see our colleagues recognised on the international stage. To have so many students involved is heartening for the future of this area of study and the role of Bristol mathematicians in shaping that future.”