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:20240626T180033Z
LOCATION:3001\, 3rd Floor
DTSTART;TZID=America/Los_Angeles:20240626T153000
DTEND;TZID=America/Los_Angeles:20240626T154500
UID:dac_DAC 2024_sess161_RESEARCH682@linklings.com
SUMMARY:Double-Win NAS: Towards Deep-to-Shallow Transformable Neural Archi
 tecture Search for Intelligent Embedded Systems
DESCRIPTION:Research Manuscript\n\nXiangzhong Luo (Nanyang Technological U
 niversity); Di Liu (Norwegian University of Science and Technology); and H
 ao Kong, Shuo Huai, and Weichen Liu (Nanyang Technological University)\n\n
 Thanks to the evolving network depth, convolutional neural networks (CNNs)
  have achieved impressive performance across various intelligent embedded 
 scenarios towards embedded intelligence. Nonetheless, this trend also lead
 s to degraded hardware efficiency as the network evolves deeper and deeper
 . In contrast, shallow networks exhibit superior hardware efficiency, whic
 h, unfortunately, suffer from inferior accuracy. To tackle this dilemma, w
 e establish the first deep-to-shallow transformable neural architecture se
 arch (NAS) paradigm, namely Double-Win NAS (DW-NAS), which is dedicated to
  automatically exploring deep-to-shallow transformable networks to marry t
 he best of both worlds. Extensive experiments on two NVIDIA Jetson intelli
 gent embedded systems clearly show the superiority of DW-NAS over previous
  state-of-the-art methods.\n\nTopic: Embedded Systems\n\nKeyword: Embedded
  Software
END:VEVENT
END:VCALENDAR