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-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:20260424T150000
DTEND;TZID=Asia/Hong_Kong:20260424T160000
DTSTAMP:20260511T071028
CREATED:20260420T022232Z
LAST-MODIFIED:20260420T022232Z
UID:115713-1777042800-1777046400@ece.hku.hk
SUMMARY:RPG Seminar – A BO-based Method for Parameter Tuning in Voltage Controlled DERs to Improve Flexibility Area at TSO-DSO Interface
DESCRIPTION:Zoom Link:\nhttps://us05web.zoom.us/j/82211644170?pwd=MUs6qBHh0EZskKbZOsacWOHAGHK4rQ.1 \nAbstract\nThe rapid proliferation of aggregated flexible resources in active distribution systems (ADSs) calls for accurate estimation of the flexibility area (FA) at the transmission system operator–distribution system operator (TSO–DSO) interface to improve operational coordination. Meanwhile\, voltage control plays a critical role in maintaining both economic and secure grid operation. However\, the relationship between voltage control parameters and the FA metric is highly nonlinear and analytically intractable\, as inverter droop coefficients and other control parameters interact with network constraints and distributed energy resource (DER) operating conditions in a coupled manner. \nTo address this challenge\, this paper proposes a Bayesian optimization (BO)-based framework that systematically tunes voltage support parameters to maximize the FA potential. The main contributions are threefold: (1) a cross-layer optimization paradigm that links inverter-level droop control with systemlevel FA characteristics\, revealing how voltage control curves shape interface power flow dynamics; (2) a reformulated FA estimation process in which the Interval-Constrained Power Flow (ICPF) model\, embedded with droop-control operations\, is cast as a Bayesian black-box problem solved efficiently via Gaussian process surrogates with a tile-coding kernel; and (3) a composite performance index\, namely the Secure Area–Cost Flexibility (SeCoF)\, which jointly quantifies the trade-off between flexibility enhancement and operational cost. Case studies on modified IEEE 13-bus and IEEE 37-bus distribution systems demonstrate the effectiveness\, robustness\, and scalability of the proposed method in improving both technical flexibility and economic performance under realistic operational constraints. \nSpeaker\nMiss Zihui HONG\nDepartment of Electrical and Computer Engineering\nThe University of Hong Kong \nBiography of the Speaker\nZihui Hong received his BEng in Electrical and Electronic Engineering from Huazhong University of Science and Technology in 2024. She is currently pursuing an MPhil degree with Prof. Hou in the Department of Electrical and Computer Engineering\, the University of Hong Kong with a research focus on Bayesian Methods for Estimation and Optimization of Flexibility Region in Power Systems. \nOrganiser\nProf. Yunhe HOU \nDepartment of Electrical and Computer Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20260424-3/
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:20260425T090000
DTEND;TZID=Asia/Hong_Kong:20260425T173000
DTSTAMP:20260511T071028
CREATED:20260310T093703Z
LAST-MODIFIED:20260326T023321Z
UID:115286-1777107600-1777138200@ece.hku.hk
SUMMARY:HKUECE: RoboLeague 2026 cum RoboCup Junior China Open (Hong Kong)
DESCRIPTION:*** Application deadline is on April 2\, 2026 (Thursday) at 8:00 pm ***\n\n🚀 Join the Excitement of Robotics!\n🌟 Please stay tuned to event-day announcements for any schedule adjustments or important updates! 🤩 \nCompetition Details\n📅 Date: April 25\, 2026 (Saturday)\n🕒 Time: 9:00 am – 5:30 pm\n📍Venue: The University of Hong Kong \nCompetition Description\nEngage your autonomous robots in thrilling challenges: \n– Soccer League (Open) （足球公開租）\n– Soccer League (Lightweight) （足球輕量組）\n– Soccer League (Standard Platform) （足球標準平台組）\n– Rescue League (Line) （搜救循線組）\n– Rescue League (Maze) （搜救迷宮組）\n– Mini-rescue League （迷你救援組） \nWho Can Join?\nTarget Audience: Secondary School Students Team\nFormation: 2-4 students ( any Forms ) per team from the same school \nKey Dates\n*Briefing Session: \n📅 Date: March 28\, 2026 (Saturday) @1:00 pm\n📍Venue: Online via Zoom \n*Application Deadline: \n📅 Date: April 2\, 2026 (Thursday) @8:00 pm \nRegistration & More Details\nhttps://linktr.ee/hkuecerl \nEnquiries\nPlease contact us via WhatsApp at 5793 9462 or email at hkuecerl@gmail.com\, using the subject line “HKUECERL2026 Enquiry”.
URL:https://ece.hku.hk/events/rl2026/
CATEGORIES:Highlights,Seminar
ATTACH;FMTTYPE=image/jpeg:https://ece.hku.hk/wp-content/uploads/2026/03/1280-3.jpg
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Hong_Kong:20260427T103000
DTEND;TZID=Asia/Hong_Kong:20260427T113000
DTSTAMP:20260511T071028
CREATED:20260422T071526Z
LAST-MODIFIED:20260422T071526Z
UID:115748-1777285800-1777289400@ece.hku.hk
SUMMARY:RPG Seminar – Toward Trustworthy Network Intrusion Detection: From Representation Learning to Multi-Agent Orchestration
DESCRIPTION:Zoom Link:\nhttps://hku.zoom.us/j/94349548162?pwd=rk4r0ND8iT1JzXjGsg293RmHKPGrjg.1 \nAbstract\nModern network intrusion detection systems face critical challenges in open and dynamic environments\, including concept drift as threat patterns evolve\, label noise as annotations are inherently unreliable\, and limited interpretability as detection decisions are difficult to understand and trust. This seminar presents our recent progress toward a trustworthy intrusion detection paradigm. We propose a representation enhancement framework that learns drift-aware and noise-robust feature embeddings\, enabling detectors to sustain high accuracy under non-stationary and imperfectly labeled conditions. We further introduce an LLM-based hierarchical multi-agent system that brings semantic reasoning into intrusion detection\, delivering both strong detection performance and human-interpretable explanations. These works chart a path from robust representation learning to next-generation agentic inference\, advancing network intrusion detection toward greater robustness\, adaptivity\, and trustworthiness. \nSpeaker\nMr Shuo YANG\nDepartment of Electrical and Computer Engineering\nThe University of Hong Kong \nBiography of the Speaker\nShuo Yang is a Ph.D. candidate in the Department of Electrical and Computer Engineering at The University of Hong Kong\, under the supervision of Prof. Edith C. H. Ngai. His current research interests include Network Security\, Trustworthy AI\, Data Mining and LLM Agent. \nOrganiser\nProf Edith C. H. NGAI \nDepartment of Electrical and Computer Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20260427-2/
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:20260427T143000
DTEND;TZID=Asia/Hong_Kong:20260427T153000
DTSTAMP:20260511T071028
CREATED:20260413T092309Z
LAST-MODIFIED:20260413T092309Z
UID:115657-1777300200-1777303800@ece.hku.hk
SUMMARY:RPG Seminar – Mobile Reasoning-as-a-Service via Distributed LLM Inference-Time Scaling
DESCRIPTION:Zoom Link:\nhttps://hku.zoom.us/j/6589092185?pwd=19EpQ4AqzgRRwQzym8vF0aaKui1J2b.1 \nAbstract\nInference-time scaling has emerged as an effective approach for enhancing the capabilities of Large Language Models (LLMs)\, addressing the growing demand for stronger reasoning without increasing model size. This novel form of LLM scaling comprises two representative approaches: explicit reasoning\, which generates intermediate chain-of-thought tokens during an explicit thinking phase\, and implicit reasoning\, which iteratively updates hidden states in the latent space without producing explicit outputs. Despite their effectiveness\, both paradigms incur substantial computational and memory overhead\, raising challenges for deployment on resource-constrained edge devices. To address these issues\, we propose a Mobile Reasoning-as-a-Service framework that treats reasoning as a computational service accessible to edge devices over wireless networks. Focusing on implicit reasoning\, we leverage its recursive structure to partition hidden-state updates between edge devices and servers\, enabling cooperative inference that allows devices to access additional cloud computation on demand. To handle dynamic wireless conditions and optimize long-term performance\, we formulate a joint computation and communication scheduling problem and solve it using a semantic Mixture-of-Experts (MoE)-based Soft Actor-Critic (SAC) algorithm to address heterogeneity in wireless conditions and task demands. Ultimately\, this work validates distributed inference-time scaling through semantic-aware collaborative reasoning services\, offering a scalable and efficient paradigm for deploying advanced LLM reasoning at the mobile edge \nSpeaker\nMr. Guanchen LIU\nDepartment of Electrical and Computer Engineering\nThe University of Hong Kong \nBiography of the Speaker\nGuanchen Liu received his B.S. degree from the Harbin Institute of Technology. He is currently an MPhil candidate in the Department of Electrical and Computer Engineering at the University of Hong Kong\, with a research focus on LLM reasoning. \nOrganiser\nProf. Kaibin HUANG \nDepartment of Electrical and Computer Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20260427/
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:20260429T160000
DTEND;TZID=Asia/Hong_Kong:20260429T170000
DTSTAMP:20260511T071028
CREATED:20260420T023103Z
LAST-MODIFIED:20260420T023132Z
UID:115717-1777478400-1777482000@ece.hku.hk
SUMMARY:RPG Seminar – Intrinsic Wavelength Tuning in InGaN/GaN MQWs Nano/MicroLEDs
DESCRIPTION:  \nAbstract\nIntrinsic wavelength tuning in InGaN/GaN multiple quantum wells (MQWs) is a critical aspect of optimizing their performance for various applications\, including display technologies and visible light communication. This project investigates the underlying mechanisms of wavelength tuning in InGaN/GaN MQWs and explores strategies to achieve precise control over the emission wavelength. Through a combination of experimental studies and theoretical modeling\, we analyze the effects of piezoelectric strain\, indium composition\, and external electric field on the emission properties of nano/microLEDs. The findings provide insights into the design and fabrication of LEDs with tuable emission characteristics\, contributing to advancements in optoelectronic device engineering. \nSpeaker\nMr. Bo LU\nDepartment of Electrical and Computer Engineering\nThe University of Hong Kong \nBiography of the Speaker\nBo LU received his BSc degree in Physics from the China University of Mining and Technology in 2018 and MSc from the Southern University of Science and Technology (HIT joint program). He is currently pursuing a PhD in the Department of Electrical and Computer Engineering at the University of Hong Kong. His research focuses on the fabrication\, characterization\, and applications of long-wavelength InGaN/GaN MQWs microLED\, particularly in wavelength tuning engineering. \nOrganiser\nProfessor Anthony Hoi Wai CHOI \nDepartment of Electrical and Computer Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20260429/
LOCATION:Room CB-603\, 6/F\, Chow Yei Ching Building\, The University of Hong Kong
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:20260505T140000
DTEND;TZID=Asia/Hong_Kong:20260505T150000
DTSTAMP:20260511T071028
CREATED:20260421T023945Z
LAST-MODIFIED:20260421T023945Z
UID:115744-1777989600-1777993200@ece.hku.hk
SUMMARY:RPG Seminar – Accuracy-Oriented Resource Scheduling for Satellite-Air-Ground Sensing\, Computing\, and Communication Systems
DESCRIPTION:  \nAbstract\nWith the evolution of 6G mobile communications\, achieving 99.999% seamless global coverage has become a key requirement. Traditional terrestrial networks still suffer from coverage blind spots in scenarios such as emergency rescue and vast ocean areas. Therefore\, building a space–air–ground integrated network by combining satellite\, aerial\, and terrestrial systems has become an important trend. This seminar focuses on mission-driven target recognition within the integrated sensing\, communication\, and computing (ISCC) framework\, emphasizing efficient end-to-end system design under physical constraints. It analyzes the theoretical lower bounds of multiclass classification error probability\, including discrimination gain and feature distribution variance\, and discusses multi-objective trade-offs and optimization frameworks. The seminar also highlights future research on user-oriented intelligent scheduling for space–air–ground collaborative resources\, with the goal of reducing end-to-end latency\, improving resource allocation efficiency\, and enhancing overall system performance. \nSpeaker\nMr Zhiyuan XU\nDepartment of Electrical and Computer Engineering\nThe University of Hong Kong \nBiography of the Speaker\nXu Zhiyuan received the B.Sc. degree from Sun Yat-sen University\, Guangzhou\, China\, in 2023. He is currently working toward the M.Phil. degree with The University of Hong Kong\, Hong Kong. His research interests include 6G communications. \nOrganiser\nProf. Kaibin HUANG \nDepartment of Electrical and Computer Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20260505/
LOCATION:Room CB-603\, 6/F\, Chow Yei Ching Building\, The University of Hong Kong
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:20260508T143000
DTEND;TZID=Asia/Hong_Kong:20260508T153000
DTSTAMP:20260511T071028
CREATED:20260430T030808Z
LAST-MODIFIED:20260430T031009Z
UID:115806-1778250600-1778254200@ece.hku.hk
SUMMARY:RPG Seminar – Novel Two-Dimensional (2D) Memory Devices: From Material Innovation to Functional Integration
DESCRIPTION:Zoom Link \nhttps://hku.zoom.us/j/92285676627?pwd=oaba8ue6daTbmrDBhxaIhYsykCedbR.1 \nAbstract\nDuring the period of the Internet of Things (IoT) and big data\, data capacity is growing exponentially\, the delay and loss caused by data transmission make the traditional von Neumann computing architecture urgently need to be overturned and restructured. This has led to a growing demand for emerging memory devices\, posing significant challenges to conventional silicon-based memory technologies. In recent years\, two-dimensional (2D) materials leveraging their unique physical properties\, ultra-thin thickness and no dangling bonds have been wide-ranging used in the fabrication of various electronic and optoelectronic memory devices. In this seminar\, we first briefly introduce several mainstream types of 2D memory devices developed in recent years and their working mechanisms. We then propose a 2D infrared-sensing memory device (ISMD) based on the Se0.3Te0.7/CuInP2S6 (CIPS) heterostructure\, where Se0.3Te0.7 serves as the channel and CIPS functions as the ferroelectric auxiliary layer. The coupling between interfacial defect trapping and ferroelectric polarization endows the device with non-volatile multi-bit memory capability programmable by electrical pulses. Meanwhile\, the device exhibits a transient infrared response at 1550 nm\, with its responsivity negatively correlated to the channel conductance. By integrating sensing\, memory\, and computing in a single device\, this work broadens the research scope of 2D memory devices. \nSpeaker\nMr. Xuyang ZHENG\nDepartment of Electrical and Computer Engineering\nThe University of Hong Kong \nBiography of the Speaker\nXuyang Zheng is an MPhil student in the Department of Electrical and Computer Engineering at The University of Hong Kong\, supervised by Prof. Can Li. He received his B.S. degree in Functional Materials from South China University of Technology (SCUT) in 2024. His research interests include 2D memory devices for neuromorphic computing. \nOrganiser\nProf. Can LI \nDepartment of Electrical and Computer Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20260508/
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:20260508T160000
DTEND;TZID=Asia/Hong_Kong:20260508T170000
DTSTAMP:20260511T071028
CREATED:20260504T020614Z
LAST-MODIFIED:20260504T020614Z
UID:115819-1778256000-1778259600@ece.hku.hk
SUMMARY:RPG Seminar – Day–Night Mechanism-Aware Causal Modeling for Wind Power Forecasting: A Physics-Guided NLSEM Framework
DESCRIPTION:Zoom Link \nhttps://hku.zoom.us/j/95263844640 \nAbstract\nWind power forecasting remains highly challenging due to the strong nonlinearity of atmospheric dynamics\,pronounced diurnal regime differences\, and substantial uncertainties in multi-source meteorological data. Conventional black-box machine-learning models mainly rely on observational correlations\, often neglecting physical constraints and causal mechanisms\, which leads to limited interpretability and poor robustness under distribution shifts. In this seminar\, we proposes a unified forecasting framework that integrates physics-constrained data construction\, multi-site causal structure learning\, and a global nonlinear structural equation model (NLSEM). The framework combines full-variable Granger causality networks with PCMCI+ to identify distinct day and night-time causal directed acyclic graphs (DAGs). Within the NLSEM\, physical monotonicity\, environmental invariance\, and counterfactual-consistency regularization are explicitly enforced. The resulting model supports causal inference through dointervention analysis\, ATE/CATE estimation\, and counterfactual reasoning. Experiments conducted on three coastal wind farms demonstrate consistent performance improvements over strong machine-learning baselines\, while revealing physically meaningful causal drivers of wind-power generation. \nSpeaker\nMr. Yuxuan WANG\nDepartment of Electrical and Computer Engineering\nThe University of Hong Kong \nBiography of the Speaker\nYuxuan Wang received the B.S. degree in New Energy Science and Engineering from Huazhong University of Science and Technology\, and the M.S. degree in electrical engineering from University of Leeds. He is currently pursuing the Ph.D. degree in electrical and electronic engineering at the Department of Electrical and Electronic Engineering\, The University of Hong Kong. His current research interests include wind power forecasting\, causal inference\, and nonlinear modeling for renewable energy systems. \nOrganiser\nProf. Yunhe HOU \nDepartment of Electrical and Computer Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20260508-2/
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:20260513T100000
DTEND;TZID=Asia/Hong_Kong:20260513T110000
DTSTAMP:20260511T071028
CREATED:20260507T082216Z
LAST-MODIFIED:20260507T082216Z
UID:115875-1778666400-1778670000@ece.hku.hk
SUMMARY:RPG Seminar – Towards Controllable and Cinematic Visual Content Generation
DESCRIPTION:Zoom Link \nhttps://hku.zoom.us/j/96029601263 \nAbstract\nThis seminar presents novel solutions to bridge the gap between creative intent and automated visual synthesis. We first address the challenge of compositional alignment in text-to-image generation. This includes T2I-CompBench(++) for systematic evaluation of multi-attribute prompts\, and GenMAC\, a multi-agent collaborative framework that enables precise semantic orchestration. Second\, we target the lack of cinematic artistry in video generation via Filmaster\, a system that injects professional camera language and rhythm into the synthesis process using real-film references. Finally\, we tackle spatio-temporal inconsistency in dynamic sequences using CineScene\, which leverages implicit 3D-aware representations to preserve scene consistency across camera trajectories. Together\, these contributions pave the way for next-generation controllable tools in digital cinematography and professional content creation. \nSpeaker\nMs Kaiyi HUANG\nDepartment of Electrical and Computer Engineering\nThe University of Hong Kong \nBiography of the Speaker\nKaiyi Huang received her BSc degree and MSc in Electrical Engineering from the Shanghai Jiao Tong University\, and a double MSc degree from Ecole Centrale Paris (joint program). She is currently pursuing a PhD in the Department of Electrical and Computer Engineering at the University of Hong Kong. Her research focuses on the image/video generation\, film generation and world model. \nOrganiser\nProf Xihui LIU \nDepartment of Electrical and Computer Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20260513/
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:20260514T103000
DTEND;TZID=Asia/Hong_Kong:20260514T113000
DTSTAMP:20260511T071028
CREATED:20260508T020416Z
LAST-MODIFIED:20260508T020416Z
UID:115879-1778754600-1778758200@ece.hku.hk
SUMMARY:RPG Seminar – HIMSA: A Heterogeneous In-Memory Computing and Searching Architecture for Efficient Attention-Based Models
DESCRIPTION:Zoom Link \nhttps://hku.zoom.us/j/99174148480?pwd=duVxaYZOJDT6MWxDh4OKOMmyo12A7A.1 \nAbstract\nThe Transformer architecture\, the foundation for modern large language models (LLMs)\, has revolutionized natural language processing and other AI domains. However\, its significant computational and memory requirements\, primarily from the matrix multiplication in the self-attention mechanism\, present major challenges for conventional hardware. While intensive research on in-memory computing (IMC) technology offers a path to overcome the memory bottleneck\, using IMC for Transformers remains challenging. This is because the dynamic matrix multiplication with frequently changing Key\, Query\, and Value matrices require frequent and costly write operations that are ill-suited for non-volatile memories (NVM) technologies like ReRAM. This work introduces HIMSA Heterogeneous In-Memory Computing and Searching Architecture\, which employs vector quantization on K and V matrices. This technique transforms the dynamic vector-matrix multiplications into static operations performed on pre-trained codebooks\, thereby eliminating the need for runtime write operations in the attention mechanism. The proposed architecture was evaluated through circuit-level simulations that account for the peripheral circuit designs\, including nearest neighbor search and the division-less Softmax operations. More importantly\, its write-free attention mechanism mitigates the concerns over limited write endurance of ReRAM devices . This work presents a promising pathway toward highly efficient NVM-based hardware acceleration for next-generation AI models. \nSpeaker\nMr Muyuan PENG\nDepartment of Electrical and Computer Engineering\nThe University of Hong Kong \nBiography of the Speaker\nPeng Muyuan received the B.S. degree at the University of Science and Technology of China majored in Applied Physics. He is currently pursuing the Ph.D. degree in electrical and electronic engineering at the Department of Electrical and Electronic Engineering\, The University of Hong Kong. His current research interests include non-volatile memory devices\, in-memory computing and related neural network accelerations. \nOrganiser\nProfessor Can LI \nDepartment of Electrical and Computer Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20260514-2/
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:20260514T133000
DTEND;TZID=Asia/Hong_Kong:20260514T143000
DTSTAMP:20260511T071028
CREATED:20260505T021413Z
LAST-MODIFIED:20260505T021413Z
UID:115830-1778765400-1778769000@ece.hku.hk
SUMMARY:RPG Seminar – Cooperative Edge AI: From Event-triggered Inference to Efficient Model Downloading
DESCRIPTION:Zoom Link \nhttp://hku.zoom.us/j/7074144117?omn=95813783034 \nAbstract\nCooperative edge AI enables edge devices and edge servers to collaboratively execute intelligent tasks under limited computation\, storage\, energy\, and communication resources. In this talk\, we discuss two complementary research directions toward communication-efficient cooperative edge AI. First\, we introduce an event-triggered cooperative inference framework for rare-event detection in edge intelligence systems. Rare events are usually infrequent but highly critical\, while conventional edge inference systems may overlook them due to data imbalance and rigid resource allocation. To address this issue\, a dual-threshold multi-exit architecture is adopted\, allowing confident normal events to be processed locally while complex or uncertain rare events are selectively offloaded to the edge server for more accurate classification. Second\, we present an efficient AI model downloading framework based on parametric-sensitivity-aware retransmission. Instead of treating all model parameters equally\, this framework exploits the unequal importance of neural network parameters and allocates wireless retransmission resources to more sensitive model packets. In this way\, downloading latency can be reduced while inference performance is preserved. The talk concludes with a discussion of future research directions in cooperative edge AI\, highlighting open challenges and opportunities in communication-efficient inference\, adaptive model deployment\, and resource-aware edge intelligence. \nSpeaker\nMr Zhou You\nDepartment of Electrical and Computer Engineering\nThe University of Hong Kong \nBiography of the Speaker\nZhou You is currently pursuing a Ph.D. degree in the Department of Electrical and Electronic Engineering at The University of Hong Kong\, under the supervision of Prof. Kaibin Huang. He received his B.Eng. degree in Electrical Engineering from the University of Wisconsin–Madison\, USA\, in 2021. His research interests include wireless communications\, edge inference\, and AI model downloading. \nOrganiser\nProf. Kaibin HUANG \nDepartment of Electrical and Computer Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20260514/
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:20260519T160000
DTEND;TZID=Asia/Hong_Kong:20260519T170000
DTSTAMP:20260511T071028
CREATED:20260420T064633Z
LAST-MODIFIED:20260420T064633Z
UID:115720-1779206400-1779210000@ece.hku.hk
SUMMARY:RPG Seminar – From Understanding to Intervention: Interpretability-Guided Methods for Improving Large Language Models
DESCRIPTION:  \nAbstract\nLarge language models have achieved impressive performance\, but improving them efficiently and reliably requires more than scaling alone. In this talk\, I present a series of works that explore how internal understanding of LLMs can be translated into practical interventions for better capability\, efficiency\, and controllability. I begin with actionable mechanistic interpretability\, introducing a unified “Locate\, Steer\, and Improve” perspective that turns model analysis into a framework for intervention. I then show how this perspective supports several concrete advances: data-free mixed-precision quantization guided by numerical and structural sensitivity\, multilingual capability enhancement through representation shifting and contrastive alignment\, personalized multi-teacher distillation that routes each prompt to its most suitable teacher\, and coarse-to-fine selective fine-tuning for mitigating catastrophic forgetting while preserving general versatility. Together\, these works reflect a common theme: interpretability is not only a tool for explaining LLMs\, but also a principled basis for designing more efficient training\, compression\, and adaptation methods. \nSpeaker\nMr Hengyuan ZHANG\nDepartment of Electrical and Computer Engineering\nThe University of Hong Kong \nBiography of the Speaker\nHengyuan Zhang is a Ph.D. candidate at the University of Hong Kong\, supervised by Prof. Ngai Wong and Prof. Hayden Kwok-Hay So. His research focuses on the improvement of efficiency and interpretability within large language models. He aims to uncover and characterize the internal processes that govern model behavior\, with the goal of improving model speciality\, interpretability\, and reliability in real-world deployments. He has published multiple papers in leading venues such as ACL\, EMNLP\, TKDD\, and NeurIPS. \nOrganiser\nProf. Ngai WONG \nDepartment of Electrical and Computer Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20260519/
LOCATION:Room CB-603\, 6/F\, Chow Yei Ching Building\, The University of Hong Kong
CATEGORIES:Seminar
ATTACH;FMTTYPE=image/jpeg:https://ece.hku.hk/wp-content/uploads/2024/11/rpg-seminar.jpg
END:VEVENT
END:VCALENDAR