Professor Julie Weeds

Professor Julie Weeds graduated from Cambridge University with a 1^st in Computer Science in 1998 and received a PhD from the University of Sussex in Natural Language Processing in 2003.  She then worked as a secondary school mathematics teacher between 2005 and 2010 before becoming a mother to twins in 2010 and returning to work part-time as a researcher at the University of Sussex in 2012.  She became a full-time lecturer in 2016, teaching and researching in the area of Natural Language Processing and Data Science more generally.  She has led projects developing conversion courses for Data Science and Artificial Intelligence and supporting students from minority backgrounds on those courses as well as projects with local companies looking to incorporate Natural Language Processing technology within their products or workflows.   She is co-director of SussexAI, a University of Sussex Centre of Excellence and she was promoted to Professor of Artificial Intelligence in 2023.

Talk Title

Have Transformers Solved Natural Language Processing?

Talk Summary

In recent years, transformer-based large language models have revolutionised Natural Language Processing (NLP).  At a recent top NLP conference, ACL-24, approximately 50% of papers included large language models in the title and this is the “go-to” technology for practically all NLP applications including text classification, information extraction, summarisation and translation.  In fact, some people would say that large language models are so good that Natural Language Processing is solved as a research area.  In this talk, I will discuss how large language models can be useful in a variety of interdisciplinary applications including work at Sussex on wildlife exploitation, propaganda and hate speech detection, and resource recommendation.  However, despite the success of these methods, there are issues around linguistic diversity, bias, explainability, trust, consistency and accountability as well as environmental sustainability. In short, as I will discuss, NLP researchers still have work to do!

This talk has been organised in partnership with the I-X Women in AI Network (IX-WAI). These events are open to everyone, regardless of gender, and we encourage all to attend for insightful discussions and workshops. 

For this seminar, IX-WAI will sponsor a networking lunch with the speaker for four women or non-binary individuals on a first-come, first-served basis. If there are no internal sign-ups, the spots will be made available to everyone. To sign up for the networking lunch, please email laburton@ic.ac.uk by midday on Tuesday 22 October. Given the reduced size of this networking lunch, please do not sign up unless you are sure you can attend, and if your circumstance change, let us know as soon as possible.