UKAIRS – a note by the programme chairs  

Our aim has been for UKAIRS to showcase the best AI research funded by UKRI whilst also providing a platform for UK AI researchers, at all career stages, to discuss emerging challenges across different fields, exchange ideas and inform the wider community (industry, government, civil society) of recent developments. The programme consists of a mix of plenary sessions, including keynote talks, lightning talks, poster and demo sessions and a community panel as well as parallel sessions of papers and panels. 

As this is the first edition of the symposium (with a full conference intended for 2026), and to help manage demand, contributions were invited only from researchers, students, and practitioners directly funded by a UKRI project or programme (including pump-priming projects and impact accelerator awards) in the last 3 years (since June 2022). Contributions from early career researchers were particularly encouraged.  

We were pleased to receive a high level of interest, and the final programme reflects the wealth of high-quality research but also the challenge of accommodating the breadth of topics within the allocated 2 days. We received 198 submissions and most work deemed relevant to UKAIRS was included in the programme as we wanted to be as inclusive as possible and showcase a wide range of research projects. In putting the programme together, we have been guided by our generous reviewer community, and particularly their assessment of the ‘fit to programme’ or likely level of interest in the research from the diverse UKAIRS audience. Work that has already been published received a very light touch review, mainly focussing on suitability and the format of presentation. Unpublished work received more rigorous review with feedback and recommendations to authors.   

The talks are organised into thematic areas which were reached through a bottom-up approach, inviting authors to include topic areas/keywords at the point of submission. Naturally dividing  a diverse set of submissions across a small number of topics (especially given space constraints) comes with a series of challenges and as a result some pieces of work are relevant to more than one of the symposium sessions, while others have relevance to important subjects that fall somewhat outside the topics that we felt were the best choice overall. 

The format of the presentation of published work already peer reviewed at international venues, as either poster or oral presentation, was determined either based on author preference or recommendation by reviewers and does not reflect the quality of the work. To further highlight papers presented as posters and published in top venues we have given them lightning talks in a plenary session. During poster sessions we also provide information about the venues where work has been published when this information was shared with us by authors. Panels were proposed by the research community and selected by experienced researchers from the UKAIRS steering committee to reflect a range of topics that are also likely to spark interesting discussions. 

To conclude we hope this first iteration of UKAIRS will be a great starting point in enabling cross-pollination between AI researchers tackling different areas and research problems, including those with societal impact. We have learnt many valuable lessons in terms of the challenges in catering for such a range of AI topics which we hope to improve on in future iterations of the conference.