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:20251127T140000
DTEND;TZID=Asia/Hong_Kong:20251127T150000
DTSTAMP:20260511T143724
CREATED:20251120T084558Z
LAST-MODIFIED:20251120T084558Z
UID:114048-1764252000-1764255600@ece.hku.hk
SUMMARY:RPG Seminar – A Continuous-Time Memristor-based Ising Solver for High-Efficiency Combinatorial Optimization
DESCRIPTION:Zoom Link: https://hku.zoom.us/j/99378601502?pwd=bKeW5GqjRbFRaQFLBmZmTBJHPSdKBf.1 \nAbstract\nSolving complex combinatorial optimization problems is a fundamental challenge that pushes conventional digital computers to their limits. While some physics-based computing approaches offer a promising alternative\, many existing systems remain trapped in a hybrid digital-analog loop\, burdened by slow\, power-hungry iterations and data conversions. \nThis work presents a fully integrated memristor-based Ising machine chip that operates as a fully analog dynamic system\, solving these problems in a single shot. Its architecture embeds the entire optimization process into the continuous physical dynamics of the circuit. By encoding the problem’s couplings as memristor conductances\, the hardware directly minimizes the system’s Hamiltonian through a single\, continuous analog transient. \nExperimental results from a 96-spin integrated chip demonstrate the system’s capability to find high-quality solutions using a quantum-inspired annealing protocol. By eliminating digital overhead entirely\, the solver achieves a nearly 10x improvement in energy efficiency and a significant speed-up. This approach opens a new avenue for creating powerful and scalable hardware accelerators for the next generation of computing. \nSpeaker\nMs. Keyi Shan\nDepartment of Electrical and Electronic Engineering\nThe University of Hong Kong \nBiography of the Speaker\nKeyi Shan is a Ph.D. student in the Department of Electrical and Electronic Engineering\, supervised by Prof. Can Li. She received her B.E. degree in Automation from Xi’an Jiaotong University\, China in 2022. Her research focuses on in-memory computing\, Ising machine\, analog computing\, combinatorial optimization\, and energy-based neural networks. \nOrganiser\nProf. Can Li\nDepartment of Electrical and Electronic Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20251127-3/
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:20251127T143000
DTEND;TZID=Asia/Hong_Kong:20251127T153000
DTSTAMP:20260511T143724
CREATED:20251120T080746Z
LAST-MODIFIED:20251120T080746Z
UID:114042-1764253800-1764257400@ece.hku.hk
SUMMARY:RPG Seminar – Toward 6G Edge AI: The Optimization and Application of Movable Antenna and Fluid Antenna
DESCRIPTION:Zoom Link:https://hku.zoom.us/j/97594921448?pwd=0AyvpTWODP87uNjZhADkvcGRrXh3V7.1 \nAbstract\nThe recently emerged movable antenna (MA) and Fluid antenna (FA) show great potential in leveraging spatial degrees of freedom for enhancing the performance of wireless systems. In future AI-embedded 6G communication networks\, MA/FA has great potential to improve the quality of service of edge AI. However\, resource allocation in MA/FA-aided systems faces unique challenges due to the non-convex and coupled constraints on antenna positions. \nIn this talk\, we will systematically reveal the challenges brought by the minimum MA/FA separation constraints at first\, and propose a penalty framework for resource allocation under such new constraints in MA/FA-aided systems. \nFurthermore\, we will also address the challenge of edge AI inference for handling the trade-off problem of model accuracy and network latency. To guarantee the high-quality of users’ service\, the latency and peak signal-to-noise ratio (PSNR) of features are considered in the objective of optimization\, and we propose an efficient algorithm under the block coordinate descent framework to solve this trade-off problem.\n \nSpeaker\nMr. Yichen Jin\nDepartment of Electrical and Electronic Engineering\nThe University of Hong Kong \nBiography of the Speaker\nYichen Jin received the B.Eng. degree from the Faculty of Automation\, Nanjing University of Science and Technology\, Nanjing\, China\, and the MSc degree from the Department of Electrical and Electronic Engineering\, The University of Hong Kong\, Hong Kong\, in 2020 and 2022\, respectively. He is currently working toward the Ph.D. degree with the Department of Electrical and Electronic Engineering\, The University of Hong Kong\, Hong Kong. His research interests include wireless communication and edge AI. \nOrganiser\nProf. Yik-Chung Wu\nDepartment of Electrical and Electronic Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20251127-2/
LOCATION:Online via Zoom
CATEGORIES:Seminar
ATTACH;FMTTYPE=image/jpeg:https://ece.hku.hk/wp-content/uploads/2024/11/rpg-seminar.jpg
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Hong_Kong:20251127T150000
DTEND;TZID=Asia/Hong_Kong:20251127T160000
DTSTAMP:20260511T143724
CREATED:20251120T032944Z
LAST-MODIFIED:20251120T032944Z
UID:114030-1764255600-1764259200@ece.hku.hk
SUMMARY:RPG Seminar – Trustworthy Tree-based Machine Learning by MoS2 Flash-based Analog CAM with Inherent Soft Boundaries
DESCRIPTION:Zoom Link: https://hku.zoom.us/j/96125660975?pwd=yg6g1tnX9xobocust8dUATRUcIan5q.1 \nAbstract\nThe rapid advancement of artificial intelligence has raised concerns regarding its trustworthiness\, especially in terms of interpretability and robustness. Tree-based models like Random Forest and XGBoost excel in interpretability and accuracy for tabular data\, but scaling them remains computationally expensive due to poor data locality and high data dependence. Previous efforts to accelerate these models with analog content addressable memory (CAM) have struggled\, due to the fact that the difficult-to-implement sharp decision boundaries are highly susceptible to device variations\, which leads to poor hardware performance and vulnerability to adversarial attacks. This work presents a novel hardware-software co-design approach using MoS2 Flash-based analog CAM with inherent soft boundaries\, enabling efficient inference with soft tree-based models. Our soft tree model inference experiments on MoS2 analog CAM arrays show this method achieves exceptional robustness against device variation and adversarial attacks while achieving state-of-the-art accuracy. Specifically\, our fabricated analog CAM arrays achieve 96% accuracy on Wisconsin Diagnostic Breast Cancer (WDBC) database\, while maintaining decision explainability. Our experimentally calibrated model validated only a 0.6% accuracy drop on the MNIST dataset under 10% device threshold variation\, compared to a 45.3% drop for traditional decision trees. This work paves the way for specialized hardware that enhances AI’s trustworthiness and efficiency. \nSpeaker\nMr. Bo Wen\nDepartment of Electrical and Electronic Engineering\nThe University of Hong Kong \nBiography of the Speaker\nBo Wen received his B.Eng. degree from the School of Materials Science and Engineering at Huazhong University of Science and Technology (HUST)\, China in 2015\, and his M.Eng. degree from the University of Chinese Academy of Sciences in 2020. He is currently pursuing a Ph.D. degree at the Department  of Electrical and Electronic Engineering under the supervision of Prof. Can Li. His research interests focus on in-memory computing\, analog content-addressable memory\, trustworthy machine learning and software-hardware co-design. \nOrganiser\nProf. Can Li\nDepartment of Electrical and Electronic Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20251127/
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