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:20260422T000711Z
LOCATION:507
DTSTART;TZID=America/Denver:20231112T160500
DTEND;TZID=America/Denver:20231112T163000
UID:submissions.supercomputing.org_SC23_sess431_ws_drbsd109@linklings.com
SUMMARY:Fast 2D Bicephalous Convolutional Autoencoder for Compressing 3D T
 ime Projection Chamber Data
DESCRIPTION:Yi Huang, Yihui Ren, Shinjae Yoo, and Jin Huang (Brookhaven Na
 tional Laboratory)\n\nHigh-energy large-scale particle colliders produce d
 ata at high speed in the order of 1 terabytes per second in nuclear physic
 s and 1 petabytes per second in high-energy physics. Time projection chamb
 er tracking detector data are usually very sparse, which presents a challe
 nge to conventional learning-free lossy compression algorithms such as SZ,
  ZFP, and MGARD.  The 3D convolutional neural network (CNN)-based approach
  named Bicephalous Convolutional Autoencoder (BCAE) outperforms traditiona
 l methods both in compression rate and in reconstruction accuracy. BCAE ca
 n also utilize the computation power of graphical processing units. Here, 
 we introduce an improved 3D CNN that achieves X% better compression ratio 
 and Y% better reconstruction accuracy measured in mean absolute error comp
 aring to BCAE.  We also introduce a novel 2D CNN variant by treating the r
 adial direction as the channel dimension, resulting a 3x in compression th
 roughput without losing too much in reconstruction accuracy.\n\nTag: Data 
 Analysis, Visualization, and Storage, Data Compression\n\nRegistration Cat
 egory: Workshop Reg Pass\n\nSession Chairs: Sheng Di (Argonne National Lab
 oratory (ANL), University of Chicago); Dingwen Tao (Institute of Computing
  Technology, Chinese Academy of Sciences; University of Chinese Academy of
  Sciences); Ana Gainaru (Oak Ridge National Laboratory (ORNL)); Jieyang Ch
 en (University of Oregon); Shadi Ibrahim (French Institute for Research in
  Computer Science and Automation (INRIA)); and Xin Liang (Oregon State Uni
 versity)\n\n
END:VEVENT
END:VCALENDAR