BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Denver
X-LIC-LOCATION:America/Denver
BEGIN:DAYLIGHT
TZOFFSETFROM:-0700
TZOFFSETTO:-0600
TZNAME:MDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0600
TZOFFSETTO:-0700
TZNAME:MST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20260422T000713Z
LOCATION:704-706
DTSTART;TZID=America/Denver:20231112T153000
DTEND;TZID=America/Denver:20231112T155000
UID:submissions.supercomputing.org_SC23_sess437_ws_shda104@linklings.com
SUMMARY:Accuracy-Constrained Efficiency Optimization and GPU Profiling of 
 CNN Inference for Detecting Drainage Crossing Locations
DESCRIPTION:Yicheng Zhang (University of California, Riverside); Dhroov Pa
 ndey (University of North Texas); Di Wu (Southern Illinois University); Tu
 rja Kundu (University of North Texas); Ruopu Li (Southern Illinois Univers
 ity); and Tong Shu (University of North Texas)\n\nThe accurate and efficie
 nt determination of hydrologic connectivity has garnered significant atten
 tion from both academic and industrial sectors due to its critical implica
 tions for environmental management. While recent studies have leveraged th
 e spatial characteristics of hydrologic features, the use of elevation mod
 els for identifying drainage paths can be influenced by flow barriers. To 
 address these challenges, our focus in this study is on detecting drainage
  crossings through the application of advanced convolutional neural networ
 ks (CNNs). In pursuit of this goal, we use neural architecture search to a
 utomatically explore CNN models for identifying drainage crossings. Our ap
 proach not only attains high accuracy (over 97% for average precision) in 
 object detection but also excels in efficiently inferring correct drainage
  crossings within a remarkably short time frame (0.268 ms). Furthermore, w
 e perform a detailed profiling of our approach on GPU systems to analyze p
 erformance bottlenecks.\n\nTag: Accelerators, Codesign, Heterogeneous Comp
 uting, Task Parallelism\n\nRegistration Category: Workshop Reg Pass\n\nSes
 sion Chairs: Tong Shu (University of North Texas), Seung-Hwan Lim (Oak Rid
 ge National Laboratory (ORNL)), Pavan Balaji (Meta AI), and Sanjukta Bhowm
 ick (University of North Texas)\n\n
END:VEVENT
END:VCALENDAR