Natural language processing

Our group focuses on various aspects of Natural Language Processing (NLP), both theoretical and applied. NLP is a subfield of Artificial Intelligence (AI) concerned on how computers deal with language, and is a growing global industry with many active research directions.

We believe NLP is interdisciplinary in nature, which is reflected in our group's interest in both NLP and domain experts that can bring the field forward, and enable impact in applications that matter.

Aims

• To understand how computers deal with language and textual content.
• To improve automatic tools dealing with language.
• To use NLP systems in applications with societal impact.

Projects

Our interdisciplinary NLP projects are funded by the European Research Council (ERC), Welsh Government, UKRI, EPSRC, industry (such as Google, Snap) and others.

A small sample of some of our projects’ outputs include:

• TweetEval: tweet classification unified benchmark + Twitter pre-trained language models
• Meemi: cross-lingual word embeddings (code+pre-trained models). Also available for Twitter
• Relation embeddings: packages to learn relation embeddings (SeVeN and RELATIVE)
• T-NER: open-source Python library for Named Entity Recognition.

Events

We organise weekly sessions on Thursdays at 13:00 (GMT).

These sessions include reading groups or talks by internal and external speakers, among others. Some sessions are open to external attendees in certain cases.

See our updated schedule of activities.