BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:America/Chicago
X-LIC-LOCATION:America/Chicago
BEGIN:DAYLIGHT
TZOFFSETFROM:-0600
TZOFFSETTO:-0500
TZNAME:CDT
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:-0500
TZOFFSETTO:-0600
TZNAME:CST
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20181221T160728Z
LOCATION:D167/174
DTSTART;TZID=America/Chicago:20181112T113000
DTEND;TZID=America/Chicago:20181112T120000
UID:submissions.supercomputing.org_SC18_sess151_ws_mlhpce105@linklings.com
SUMMARY:Automated Labeling of Electron Microscopy Images Using Deep Learni
 ng
DESCRIPTION:Workshop\nDeep Learning, Machine Learning, Workshop Reg Pass\n
 \nAutomated Labeling of Electron Microscopy Images Using Deep Learning\n\n
 Weber, Ophus, Ramakrishnan\n\nSearching for scientific data requires metad
 ata providing a relevant context. Today, generating metadata is a time and
  labor intensive manual process that is often neglected, and important dat
 asets are not accessible through search. We investigate the use of machine
  learning to generalize metadata from a subset of labeled data, thus incre
 asing the availability of meaningful metadata for search. Specifically, we
  consider electron microscopy images collected at the National Center for 
 Electron Microscopy at the Lawrence Berkeley National Laboratory and use o
 f deep learning to discern characteristics from a small subset of labeled 
 images and transfer labels to the entire image corpus.\n\nRelatively small
  training set sizes and a minimum resolution of 512x512 pixels required by
  the application domain pose unique challenges. We overcome these challeng
 es by using a simple yet powerful convolutional network architecture that 
 limits the number of free parameters to lower the required amount of compu
 tational power and reduce the risk of overfitting. We achieve a classifica
 tion accuracy of approximately 80% in discerning between images recorded i
 n two operating modes of the electron microscope---transmission electron m
 icroscopy (TEM) and scanning transmission electron microscopy (STEM). We u
 se transfer learning–i.e., re-using the pre-trained convolution layers fro
 m the TEM vs. STEM classification problem–to generalize labels and achieve
  an accuracy of approximately 70% despite current experiments being limite
 d to small training data sets. We present these predictions as suggestions
  to domain scientists to accelerate the labeling process with the goal of 
 further validating our approach and improving the accuracy of automaticall
 y created labels.
URL:https://sc18.supercomputing.org/presentation/?id=ws_mlhpce105&sess=ses
 s151
END:VEVENT
END:VCALENDAR