BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Department of Electrical and Computer Engineering (HKUECE) 電機與計算機工程系 - ECPv6.15.20//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Department of Electrical and Computer Engineering (HKUECE) 電機與計算機工程系
X-ORIGINAL-URL:https://ece.hku.hk
X-WR-CALDESC:Events for Department of Electrical and Computer Engineering (HKUECE) 電機與計算機工程系
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:Asia/Hong_Kong
BEGIN:STANDARD
TZOFFSETFROM:+0800
TZOFFSETTO:+0800
TZNAME:HKT
DTSTART:20250101T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Asia/Hong_Kong:20260106T110000
DTEND;TZID=Asia/Hong_Kong:20260106T120000
DTSTAMP:20260510T191824
CREATED:20251205T070843Z
LAST-MODIFIED:20251205T070843Z
UID:114342-1767697200-1767700800@ece.hku.hk
SUMMARY:Seminar on Machine Learning\, Artificial Intelligence\, Neuro Imaging Focusing on Longevity and Dementia (MANIFOLD)
DESCRIPTION:Abstract\nBrain health is one of the key societal challenges for the 21st century\, and much progress has been made in understanding and treating brain health conditions\, aided by growing use of neuroimaging. Meanwhile\, artificial intelligence and machine learning (AI/ML) technologies have revolutionised many domains\, including healthcare. However\, there is still a translational gap between AI/ML methods and the use of neuroimaging to detect\, treat and care for people with neurodegenerative or neurodevelopmental conditions. My talk will provide an overview of the research of the MANIFOLD lab at UCL\, that aims to bridge this gap and provide clinically useful neuroimaging tools to improve brain health. I will focus on methods that emphasise the individual patient\, namely the brain-age paradigm and neuroanatomical normative modelling\, applied to Alzheimer’s disease and dementia with Lewy bodies and frontotemporal dementia. Beyond this\, I will talk about our research in explainable AI (XAI)\, AI/ML data fusion\, automated ML and accessible MRI using portable scanners and how we have or plan to apply these in studies of brain diseases. \nSpeaker\nProf. James COLE\nProfessor of Neuroimage Computing\,\nUCL Hawkes Institute and the Dementia Research Centre (DRC)\,\nUniversity College London (UCL) \nSpeaker’s Biography\nJames Cole is Professor of Neuroimage Computing at the UCL Hawkes Institute and the Dementia Research Centre (DRC) at University College London (UCL). His research interests include brain ageing\, neurological and psychiatric diseases\, with a particular focus on ageing\, neurodegeneration and dementia. His work uses machine learning\, deep learning and related statistical methods with the goal of developing clinically useful neuroimaging tools. He is Principal Investigator of the MANIFOLD Lab. \nOrganiser\nProf. Ed Xuekui WU\nChair of Biomedical Engineering\,\nLam Woo Professorship in Biomedical Engineering\nDepartment of Electrical and Electronic Engineering\,\nThe University of Hong Kong \nAcknowledgement\nTam Wing Fan Innovation Wing Two\n\nAll are welcome!
URL:https://ece.hku.hk/events/20260106-2/
LOCATION:Tam Wing Fan Innovation Wing Two\, G/F\, Run Run Shaw Building\, The University of Hong Kong
CATEGORIES:Highlights,Seminar
ATTACH;FMTTYPE=image/jpeg:https://ece.hku.hk/wp-content/uploads/2025/12/1280-1.jpg
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Hong_Kong:20260106T143000
DTEND;TZID=Asia/Hong_Kong:20260106T153000
DTSTAMP:20260510T191824
CREATED:20251204T020220Z
LAST-MODIFIED:20251204T020220Z
UID:114335-1767709800-1767713400@ece.hku.hk
SUMMARY:Seminar on Automatic Rank Determination for Low-Rank Adaptation via Submodular Function Maximisation
DESCRIPTION:Abstract\nIn this talk\, we will introduce SubLoRA\, a rank determination method for Low-Rank Adaptation (LoRA) based on submodular function maximisation. In contrast to prior approaches\, such as AdaLoRA\, that rely on first-order (linearised) approximations of the loss function\, SubLoRA utilises second-order information to capture the potentially complex loss landscape by incorporating the Hessian matrix. We show that the linearization becomes inaccurate and ill-conditioned when the LoRA parameters have been well optimised\, motivating the need for a more reliable and nuanced second-order formulation. To this end\, we reformulate the rank determination problem as a combinatorial optimisation problem with a quadratic objective. However\, solving this problem exactly is NP-hard in general. To overcome the computational challenge\, we introduce a submodular function maximisation framework and devise a greedy algorithm with approximation guarantees. We derive a sufficient and necessary condition under which the rank-determination objective becomes submodular\, and construct a closed-form projection of the Hessian matrix that satisfies this condition while maintaining computational efficiency. Our method combines solid theoretical foundations\, second-order accuracy\, and practical computational efficiency. We further extend SubLoRA to a joint optimisation setting\, alternating between LoRA parameter updates and rank determination under a rank budget constraint. Extensive experiments on fine-tuning physics-informed neural networks (PINNs) for solving partial differential equations (PDEs) demonstrate the effectiveness of our approach. Results show that SubLoRA outperforms existing methods in both rank determination and joint training performance. \nSpeaker\nDr. Yihang GAO\nDepartment of Mathematics\,\nNational University of Singapore (NUS)\, Singapore \nSpeaker’s Biography\nYihang GAO is currently a Research Fellow in the Department of Mathematics at the National University of Singapore (NUS)\, Singapore. He received the B.S. degree in Mathematics and Applied Mathematics from Zhejiang University\, China\, in 2020\, and the Ph.D. degree in Mathematics from The University of Hong Kong (HKU)\, Hong Kong SAR\, in 2024. His research interests include mathematical machine learning\, optimisation\, and scientific computing. \nOrganiser\nProf. Kaibin HUANG\nDepartment of Electrical and Electronic Engineering\,\nThe University of Hong Kong \nAll are welcome!
URL:https://ece.hku.hk/events/20260106-1/
LOCATION:Room CB-603\, 6/F\, Chow Yei Ching Building\, The University of Hong Kong
CATEGORIES:Highlights,Seminar
ATTACH;FMTTYPE=image/jpeg:https://ece.hku.hk/wp-content/uploads/2025/12/1280.jpg
END:VEVENT
END:VCALENDAR