Human-Centred Agentic Science: Sensemaking, Provenance, and Human-AI Collaboration

Speaker: Dr Kai Xu
University of Nottingham

Title: Human-Centred Agentic Science: Sensemaking, Provenance, and Human-AI Collaboration

Date: Monday, 4 May 2026

Time: 4:00pm to 5:00pm

Venue: Lecture Theater F
(Leung Yat Sing Lecture Theater), near lift 25/26, HKUST

Abstract:

The performance of machine learning models, particularly large language models and generative AI, has dramatically improved in recent years. Despite these advancements, many tasks requiring specialised knowledge and analytical reasoning remain beyond the reach of full automation. I believe these tasks still require a collaborative approach that combines human expertise with machine intelligence, which is known as human-in-the-loop or mixed-initiative. In this talk, I will showcase some of my past and current projects, illustrating how sensemaking and provenance can help tackle this challenge and my latest interest in agentic science also fits into this through human-AI collaboration.

Biography:

Dr Kai Xu is an Associate Professor in the School of Computer Science at the University of Nottingham, which is part of the Russell Group that includes top UK universities such as Cambridge and Oxford. He is the co-director of its Visualization group (VisTAG) and founder of the Nottingham special interest group on Generative AI (GAIN). In addition, he is the chair of the UK Branch of EuroGraphics, the largest European association of Computer Graphics professionals (sponsor of EuroVis conference), and a co-organiser of the Visualisation group at the Alan Turing Institute, UK's national centre for AI and Data Science. His current research focus is human-centred machine learning, particularly human-AI teaming, that enables the collaboration between users and generative AI systems. His research has been funded by UK and European research councils, government department (such as Ministry of Defence), and large multinational companies (such as CGI and Genetec), with a total project budget of over £12 million. He has published extensively in the top visualisation and HCI journals and conferences, and some of his work was recognised with international awards.