Maria Liakata: AdSoLve: Addressing Sociotechnical limitations of LLMs for Medical & Social Computing

In this talk, Professor Maria Liakata, Professor of Natural Language Processing at Queen Mary University of London, delves into the socio-technical limitations of large language models (LLMs) in medical and social computing. Co-leading the NLP and Data Science for Mental Health interest groups at the Alan Turing Institute, Professor Liakata discusses her team’s groundbreaking project, funded by UKRI, focused on creating new evaluation benchmarks and methodologies to address biases, privacy concerns, and explainability in LLMs for healthcare and social contexts. She highlights key project goals, such as improving temporal reasoning, reducing biases, and enhancing the usability of LLMs in critical domains. This multidisciplinary initiative brings together experts from various fields, including healthcare, ethics, and law, to co-create solutions with industry partners. Watch this talk for insights into the future of AI in healthcare and social computing.