Events

Date: 3 June 2026   Time: 16:00 - 17:00     

Location: David Sizer Lecture Theatre, Bancroft Building, Mile End, E1 4NS Map 

Abstract:
Large language models contain rich and diverse knowledge, giving them the potential to perform well on tasks far beyond language. However, directly applying them to non-linguistic tasks such as human behavior analysis or motion modeling is often challenging, because the inputs for these tasks (e.g., videos or skeletal sequences) are not naturally compatible with the desired input of language models. This talk introduces a "linguistic" perspective on addressing this issue: transforming such non-linguistic signals into natural language–like "sentences," thereby enabling large language models to operate more effectively on these tasks.

Bio:
Jun Liu is Professor (Chair in Digital Health) at the School of Computing and Communications, Lancaster University. He earned his PhD from Nanyang Technological University in 2019, subsequently serving as faculty at Singapore University of Technology and Design from 2019 to 2024. Prior to his academic career, he worked at Tencent from 2014 to 2015. He is a Fellow of BCS. He obtained the Best Paper Awards from PREMIA (2016, 2019), the Best Doctoral Thesis Award from EEE at NTU (2020), and the IEEE VSPC Rising Star Honorable Mention Award (2024). He was nominated for the Singapore President's Young Scientist Award in 2024. He serves as Associate Editor-in-Chief for Pattern Recognition, Senior Area Editor for IEEE Transactions on Image Processing, and Associate Editor for several leading journals. He is General Chair of BMVC 2026 and Program Chair of BMVC 2025, and has served as Area Chair for premier AI and machine learning conferences. His research focuses on AI, computer vision, machine learning, and digital health applications.