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:20240626T180034Z
LOCATION:3010\, 3rd Floor
DTSTART;TZID=America/Los_Angeles:20240625T154500
DTEND;TZID=America/Los_Angeles:20240625T160000
UID:dac_DAC 2024_sess127_RESEARCH626@linklings.com
SUMMARY:Lightator: An Optical Near-Sensor Accelerator with Compressive Acq
 uisition Enabling Versatile Image Processing
DESCRIPTION:Research Manuscript\n\nMehrdad Morsali (New Jersey Institute o
 f Technology); Brendan Reidy (University of South Carolina); Deniz Najafi 
 (New Jersey Institute of Technology); Sepehr Tabrizchi (University of Nebr
 aska, Lincoln); Mohsen Imani (University of California, Irvine); Mahdi Nik
 dast (Colorado State University); Arman Roohi (University of Nebraska, Lin
 coln); Ramtin Zand (University of South Carolina); and Shaahin Angizi (New
  Jersey Institute of Technology)\n\nThis paper proposes a high-performance
  and energy-efficient optical near-sensor accelerator for vision applicati
 ons, called Lightator. Harnessing the promising efficiency offered by phot
 onic devices, Lightator features innovative compressive acquisition of inp
 ut frames and fine-grained convolution operations for low-power and versat
 ile image processing at the edge for the first time. This will substantial
 ly diminish the energy consumption and latency of conversion, transmission
 , and processing within the established cloud-centric architecture as well
  as recently designed edge accelerators. Our device-to-architecture simula
 tion results show that with favorable accuracy, Lightator achieves 84.4 Ki
 lo FPS/W and reduces power consumption by a factor of ~24x and 73x on aver
 age compared with existing photonic accelerators and GPU baseline\n\nTopic
 : Design\n\nKeyword: Emerging Models of Computation\n\nSession Chairs: Sud
 hakar Pamarti (UCLA) and Qingxue Zhang (Purdue University)
END:VEVENT
END:VCALENDAR