BEGIN:VCALENDAR
VERSION:2.0
PRODID:Linklings LLC
BEGIN:VTIMEZONE
TZID:Europe/Stockholm
X-LIC-LOCATION:Europe/Stockholm
BEGIN:DAYLIGHT
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
TZNAME:CEST
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=-1SU
END:DAYLIGHT
BEGIN:STANDARD
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
TZNAME:CET
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=10;BYDAY=-1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTAMP:20190719T085744Z
LOCATION:HG E 3
DTSTART;TZID=Europe/Stockholm:20190614T103000
DTEND;TZID=Europe/Stockholm:20190614T110000
UID:submissions.pasc-conference.org_PASC19_sess165_msa301@linklings.com
SUMMARY:Autoencoders for Anomaly Detection in HEP
DESCRIPTION:Minisymposium\nComputer Science and Applied Mathematics, Physi
 cs\n\nAutoencoders for Anomaly Detection in HEP\n\nThompson\n\nMachine lea
 rning methods are being increasingly and successfully applied to many diff
 erent physics problems. In this talk I will address how autoencoders can b
 e employed in model-independent, data-driven searches for new physics at t
 he LHC. Autoencoders are an unsupervised machine learning method, which I 
 will show can be applied directly to a  background-dominated signal r
 egion during both the training and testing  phase. I will then show h
 ow the autoencoder could be used in a search for a resonance in the QCD fa
 t jet-mass spectrum. This search makes use of an adversary, which decorrel
 ates the autoencoder output from the jet-mass and  therefore enables 
 us to perform a bump-hunt directly on data.
END:VEVENT
END:VCALENDAR