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:20240101T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Asia/Hong_Kong:20251128T110000
DTEND;TZID=Asia/Hong_Kong:20251128T120000
DTSTAMP:20260511T143726
CREATED:20251111T032230Z
LAST-MODIFIED:20251111T041157Z
UID:113860-1764327600-1764331200@ece.hku.hk
SUMMARY:RPG Seminar – Lightweight Learning for the Coordination of Distributed Energy Resources
DESCRIPTION:Zoom Link: https://hku.zoom.us/j/8957840635?pwd=jB4IyfmX0hTbEjn9W0LVEs31VhDw0e.1&omn=97635631185 \nAbstract\nThe proliferation of distributed energy resources presents significant coordination challenges due to their scale and variability. While traditional centralized methods are hindered by high communication and computational costs\, resource-constrained edge devices struggle with conventional algorithms. This paper aims to bridge this gap by developing lightweight learning approaches for edge devices\, enabling scalable and efficient coordination of distributed resources. The work focuses on three key analyses: descriptive analysis (non-intrusive load monitoring)\, predictive analysis (load forecasting)\, and prescriptive analysis (energy management for market participation). Ultimately\, these lightweight algorithms are implemented on established hardware testbeds\, paving the way for low-cost\, high-efficiency coordination of massive\, distributed assets. \n  \nSpeaker\nMr. Yehui LI\nDepartment of Electrical and Electronic Engineering\nThe University of Hong Kong \nBiography of the Speaker\nYehui Li received the B.S. degree in electronic science and technology from Harbin Institute of Technology in 2022. He is currently pursuing the Ph.D. degree in electrical and electronic engineering with the University of Hong Kong. His current research interests include data analytics and edge intelligence in smart grids. \nOrganiser\nProf. Yi Wang\nDepartment of Electrical and Electronic Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20251128-1/
LOCATION:Online via Zoom
CATEGORIES:Seminar
ATTACH;FMTTYPE=image/jpeg:https://ece.hku.hk/wp-content/uploads/2024/11/rpg-seminar.jpg
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Hong_Kong:20251128T110000
DTEND;TZID=Asia/Hong_Kong:20251128T120000
DTSTAMP:20260511T143726
CREATED:20251121T023324Z
LAST-MODIFIED:20251121T023324Z
UID:114064-1764327600-1764331200@ece.hku.hk
SUMMARY:RPG Seminar – Brain-Inspired Structural Optimization: Edge Pruning and Kernel Pruning Across Analog and Digital RRAM-Based Compute-in-Memory.
DESCRIPTION:Zoom Link: https://hku.zoom.us/j/93194207095?pwd=se5Jt0b8jIM7nz3yy9YZdNrWJIm818.1 \nAbstract\nThis seminar introduces two complementary pruning strategies implemented directly on RRAM-based compute-in-memory hardware. The first approach uses the intrinsic randomness of analog RRAM electroforming to build an over-parameterized random-weight network\, where edge pruning selects an efficient sub-network without requiring precise conductance tuning. This enables robust topology optimization while minimizing programming complexity.\nThe second approach is realized on a fully digital reconfigurable RRAM logic architecture\, where in-memory XOR/AND operations measure kernel similarity and dynamically prune redundant convolution kernels during training. Together\, these two pruning mechanisms illustrate a unified hardware–algorithm co-design philosophy: pruning is not a post-processing step\, but a native in-memory operation that co-optimizes connectivity\, computation\, and resource efficiency. This synergy highlights a scalable path toward adaptive\, energy-efficient RRAM-based AI accelerators. \nSpeaker\nMr. Songqi Wang\nDepartment of Electrical and Electronic Engineering\nThe University of Hong Kong \nBiography of the Speaker\nSongqi Wang received his B.Sc. degree from Huazhong University of Science and Technology\, and is currently a fourth-year Ph.D. candidate in the Department of Electrical and Electronic Engineering at The University of Hong Kong under the supervision of Prof. Han Wang. His research interests mainly include RRAM-based compute-in-memory architectures\, secure and intelligent edge-computing systems\, and software–hardware co-design for differential-equation-based models. \nOrganiser\nProf. Han Wang\nDepartment of Electrical and Electronic Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20251128/
LOCATION:Online via Zoom
CATEGORIES:Seminar
ATTACH;FMTTYPE=image/jpeg:https://ece.hku.hk/wp-content/uploads/2024/11/rpg-seminar.jpg
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Hong_Kong:20251128T140000
DTEND;TZID=Asia/Hong_Kong:20251128T150000
DTSTAMP:20260511T143726
CREATED:20251124T040333Z
LAST-MODIFIED:20251124T040333Z
UID:114204-1764338400-1764342000@ece.hku.hk
SUMMARY:RPG Seminar – Tackling Instability and Redundancy in Diffusion-Based Generative Models
DESCRIPTION:Zoom Link: https://hku.zoom.us/j/99109748447?pwd=JSHjhMjma2hylEHbOCcLr3fJRCOoJq.1 \nAbstract\nThis seminar presents novel solutions to tackle instability and redundancy in modern generative models. We first address the high-variance optimization challenges in Conditional Flow Matching (CFM) by introducing the Stable Velocity framework. This includes StableVM for robust training stability and StableVS\, a finetuning-free accelerator that doubles sampling speed. Second\, we target spatial redundancy in super-resolution via the Quadtree Diffusion Model (QDM). QDM utilizes a quadtree-guided masking strategy to focus computation solely on information-rich regions. Together\, these contributions pave the way for more stable\, efficient\, and scalable generative models. \nSpeaker\nMr. Donglin Yang\nDepartment of Electrical and Electronic Engineering\nThe University of Hong Kong \nBiography of the Speaker\nDonglin Yang is an MPhil student in the Department of Electrical and Electronic Engineering\, supervised by Prof. Xiaojuan Qi. He received his B.Eng. degree from Tsinghua University. His current research focuses on deep generative models\, with a particular emphasis on theoretical optimization for diffusion and flow-based models. \nOrganiser\nProf. Xiaojuan Qi\nDepartment of Electrical and Electronic Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20251128-3/
LOCATION:Online via Zoom
CATEGORIES:Seminar
ATTACH;FMTTYPE=image/jpeg:https://ece.hku.hk/wp-content/uploads/2024/11/rpg-seminar.jpg
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Hong_Kong:20251128T150000
DTEND;TZID=Asia/Hong_Kong:20251128T160000
DTSTAMP:20260511T143726
CREATED:20251112T081230Z
LAST-MODIFIED:20251112T081230Z
UID:113875-1764342000-1764345600@ece.hku.hk
SUMMARY:RPG Seminar – Collaborative Load Forecasting via Multi-Party Data Sharing
DESCRIPTION:Zoom Link: https://hku.zoom.us/j/98873959228 \nAbstract\nAccurate load forecasting is fundamental to the stability and efficiency of modern power grids. While collaborative approaches that leverage multi-party data sharing can significantly enhance forecasting accuracy\, they also introduce complex challenges. Effective collaboration is often hindered by data heterogeneity across participants\, critical data privacy concerns\, and the lack of clear incentives for sharing. This seminar aims to bridge this gap by presenting a comprehensive framework for collaborative load forecasting via multi-party data sharing. The work focuses on three key areas: first\, handling data heterogeneity through personalization strategies; second\, enhancing data privacy with distributed learning techniques; and third\, fostering collaboration through an incentive-driven model trading mechanism. Ultimately\, this framework paves the way for a secure\, efficient\, and economically viable ecosystem for multi-party collaboration\, enabling more intelligent load forecasting paradigm. \nSpeaker\nMr. Dalin Qin\nDepartment of Electrical and Electronic Engineering\nThe University of Hong Kong \nBiography of the Speaker\nDalin Qin received the B.S. degree in electrical engineering and its automation from South China University of Technology in 2022. He is currently pursuing the Ph.D. degree in electrical and electronic engineering at the University of Hong Kong. His current research interests include data analytics and data sharing in smart grids. \nOrganiser\nProf. Yi Wang\nDepartment of Electrical and Electronic Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20251128-2/
LOCATION:Online via Zoom
CATEGORIES:Seminar
ATTACH;FMTTYPE=image/jpeg:https://ece.hku.hk/wp-content/uploads/2024/11/rpg-seminar.jpg
END:VEVENT
END:VCALENDAR