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:20240101T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Asia/Hong_Kong:20251120T150000
DTEND;TZID=Asia/Hong_Kong:20251120T160000
DTSTAMP:20260511T154005
CREATED:20251113T070933Z
LAST-MODIFIED:20251113T070933Z
UID:113894-1763650800-1763654400@ece.hku.hk
SUMMARY:RPG Seminar – Fast and efficient genomic analysis with memristor-based in-memory computing hardware
DESCRIPTION:Zoom Link: https://hku.zoom.us/j/99194942035?pwd=W2Dch8eGCZCClFk8j97JlRvpcTrpMP.1 \nAbstract\nAdvances in third-generation sequencing (TGS) have unlocked the potential for portable\, real-time genomic analysis\, but data processing remains a critical bottleneck hindering practical on-site applications. The massive\, error-prone data streams generated by these sequencers overwhelm traditional von Neumann architectures\, which are limited by costly data movement. This presentation introduces two novel in-memory computing (IMC) hardware-software codesigns developed to accelerate genomic analysis directly in memory\, specifically targeting the challenges of high error rates and raw signal data. \nThe first work\, ShiftCAM\, addresses the high insertion and deletion (indel) error rates in basecalled reads\, a key challenge for existing CAM-based accelerators. ShiftCAM is a novel time-domain Content Addressable Memory (CAM) that efficiently calculates the Shifted Hamming Distance to better approximate the computationally expensive edit distance. This approach\, combined with a hardware-specific “Modification to Accidental Match” strategy\, significantly reduces false positives. Simulations demonstrate that ShiftCAM achieves a 2.1× higher F1 score in contamination analysis and offers a 29.5× speedup and 9.4× higher energy efficiency over state-of-the-art in-memory classifiers. \nThe second work presents a memristor-based codesign that bypasses basecalling entirely to process raw\, analog sequencer signals directly in analog memory. This system merges the traditionally separate steps of basecalling and read mapping. By exploiting intrinsic memristor device noise for locality-sensitive hashing and implementing parallel approximate search\, our fully integrated chip experimentally demonstrates high-accuracy (97.15% F1 score) infectious disease detection from raw signals. This direct-processing approach yields a 51× speed-up and 477× energy saving over a conventional ASIC. \nCollectively\, these two works demonstrate that specialized in-memory computing architectures provide a powerful and viable solution for integration with portable sequencers. By tackling bottlenecks from indel-rich reads (ShiftCAM) to raw analog signals (memristor-codesign)\, these approaches pave the way for true real-time\, on-site genomic analysis in fields like personalized medicine and metagenomics. \nSpeaker\nMr. Peiyi He\nDepartment of Electrical and Electronic Engineering\nThe University of Hong Kong \nBiography of the Speaker\nHe\, Peiyi received B.E degree from School of Integrated Circuits\, Tsinghua University\, Beijing\, China\, in 2023. He is currently pursuing the Ph.D. degree with the Department of Electrical and Electronic Engineering under the supervision of Prof. Can Li. His research interests mainly include in-memory computing\, content-addressable memory\, analog computing\, bioinformatics and computational biology. \nOrganiser\nProf. Can Li\nDepartment of Electrical and Electronic Engineering\, The University of Hong Kong \nAll are welcome.
URL:https://ece.hku.hk/events/20251120-2/
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:20251120T160000
DTEND;TZID=Asia/Hong_Kong:20251120T170000
DTSTAMP:20260511T154005
CREATED:20251104T030128Z
LAST-MODIFIED:20251104T030338Z
UID:113822-1763654400-1763658000@ece.hku.hk
SUMMARY:Exploring Careers in Industry: Quantitative Research Talk
DESCRIPTION:About the Talk\nThe talk is co-organised by Susquehanna and Prof. Kenneth Kin-Yip WONG from the Department of Electrical and Electronic Engineering at The University of Hong Kong. This is a unique opportunity to explore careers in quant trading\, hear firsthand from an experienced researcher\, and connect with industry professionals. \nThe speaker\, Dr. Davor OBRADOVIC\, holds a PhD in Computer Science from the University of Pennsylvania and has been a Quantitative Researcher at Susquehanna for 24 years. He’ll share insights into the quant trading landscape\, how academic research translates into solving complex trading problems\, and what life is like at Susquehanna. \nWhy Attend?\n\nDiscover how your academic background can thrive in industry\nGain insider knowledge about the quant trading field\nNetwork with Susquehanna professionals over refreshments\nReceive exclusive Susquehanna-branded merchandise\n\nRefreshments will be provided during the talk. 😊\n \nTarget Audience\nEEE RPg Students and Postdocs are welcome! \nRegister Now\nhttps://ece.hku.hk/20251120-s \nWe look forward to seeing you at the talk!
URL:https://ece.hku.hk/events/20251120-1/
LOCATION:Room LE-9\, LG2/F\, Library Extension Building (LE)\, The University of Hong Kong
CATEGORIES:Career Talks,Highlights,Seminar
ATTACH;FMTTYPE=image/jpeg:https://ece.hku.hk/wp-content/uploads/2025/11/1920.jpg
END:VEVENT
END:VCALENDAR