BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Los_Angeles
X-LIC-LOCATION:America/Los_Angeles
BEGIN:DAYLIGHT
TZOFFSETFROM:-0800
TZOFFSETTO:-0700
TZNAME:PDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0700
TZOFFSETTO:-0800
TZNAME:PST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260402T024533Z
LOCATION:3008\, Level 3
DTSTART;TZID=America/Los_Angeles:20250624T153000
DTEND;TZID=America/Los_Angeles:20250624T154500
UID:dac_DAC 2025_sess157_RESEARCH2178@linklings.com
SUMMARY:SSpMV: A Sparsity-aware SpMV Framework Empowered by Multimodal Mac
 hine Learning
DESCRIPTION:Shengle Lin, Chubo Liu, and Yan Ding (Hunan University); Joey 
 Zhou (Agency for Science, Technology and Research (A*STAR)); and Kenli Li,
  Wangdong Yang, and Xianhao He (Hunan University)\n\nSparse Matrix-Vector 
 Multiplication (SpMV) is an essential sparse operation in scientific compu
 ting and artificial intelligence.\nEfficiently adapting SpMV algorithms to
  diverse matrices and architectures requires a framework capable of accura
 tely recognizing sparse patterns and selecting the optimal implementation.
 \nIn this work, we introduce Sparsity-aware SpMV (SSpMV), a framework that
  integrates expert-designed features with multimodal representations to ad
 aptively predict the best-performing algorithm and parameters.\nFor this p
 urpose, we design a multimodal neural network called MM-Adapter, to captur
 e diverse modalities to represent the computational features of SpMV.\nExp
 erimental results demonstrate that MM-Adapter achieves the highest accurac
 y of 81.05%, outperforming existing SpMV prediction models.\nFurthermore, 
 SSpMV consistently delivers substantial performance improvements over stat
 e-of-the-art sparse libraries across various multi-core platforms.\n\nTopi
 cs: Systems\n\nTracks: SYS3: Embedded Software\n\nSession Chairs: Hoeseok 
 Yang (Santa Clara University) and Younghyun Kim (Purdue University)\n\n
END:VEVENT
END:VCALENDAR