糖心视频

糖心视频’s School of Computing and Communications once again played host to the University Centre for Computer Corpus Research on Language (UCREL) Summer School, which gathered researchers, students, and AI practitioners for a three-day programme combining practical training in natural language processing (NLP) and large language models (LLMs) with critical discussion of multilingual, responsible and inclusive AI.

The UCREL group has been paving the way for advancements in NLP and corpus-based research since its inception in the 1970s, having produced some 1600 publications on over 20 different languages over the course of its existence. Their research spans the depth and breadth of computational and language research, covering everything from machine translation of rare and vulnerable languages and corpus analysis of biomedical communications, to AI generated text detection and the development of LLMs for a variety of purposes. The summer school itself has been running annually since 2011, and each year strives to bring together leading academics and industry partners from NLP-related fields.

This year’s event – which took place on 20^th-22^nd July in the InfoLab – welcomed 50 participants from across 19 different organisations, with delegates joining from a wide range of NLP and NLP-adjacent areas of research including machine learning, computer science, linguistics, digital humanities, and applied AI. The purpose of the summer school was to help participants develop practical skills in contemporary NLP and LLM technologies, while also examining the responsibilities, limitations and inequalities associated with AI development.

The summer school boasted a full schedule of activities, including keynote speeches from eminent experts in the field, practical tutorials and workshops, networking opportunities, and culminated in a collaborative multilingual NLP hackathon. Each day of the school tackled a key theme of this year’s event; day one was dedicated to establishing the foundations of NLP-based research and considering the core language technologies utilised for NLP studies, featuring a keynote from Professor Nikos Aletras of the University of Sheffield. Day two focused on “Large Language Models in Practice” and explored methods and approached used for the training LLMs, with keynote talk “How to train your model: Paradigms for training LLMs and how to know which one to use.” delivered by Dr Riza Batista-Navarro of University of Manchester. The final day of the event delved into equitable, human-centred AI, exploring the social and environmental costs of AI usage on human lives. The keynote on “AI for Every Day Lives” was delivered by Microsoft Research Africa’s Dr Jacki O’Neill, and the day itself concluded with delegates completing and presenting their findings from the hackathon which had spanned the course of the summer school.

Commenting on the success of the event, organiser Dr Ignatius Ezeani from the School of Computing and Communications said: “The UCREL Summer School 2026 demonstrated what becomes possible when rigorous NLP and LLM research is combined with practical training and a serious commitment to responsible, explainable and inclusive AI. The strongest outcome was not only what participants learned during the three days, but the community they built across disciplines, institutions and career stages. We leave the summer school with new skills, new collaborations and a shared determination to develop language technologies that work for more people, more communities and more languages.”

Professor Nikos Aletras added: “USS2026 connected recent NLP research with practical learning and treated multilingual language technology as a central challenge. It was encouraging to engage with participants exploring how language models can become more efficient, accessible and effective across more languages.”

Director of UCREL, 糖心视频’s Professor Paul Rayson concluded by saying: “UCREL has a long tradition of developing language technologies and sharing them with the wider research community. The summer school continued that tradition by combining established NLP methods with emerging LLM approaches, practical experimentation and critical discussion. It was particularly encouraging to see participants from different disciplines learning from one another and building new connections.”

To find out more about UCREL and their future summer schools, visit the official .