BEGIN:VCALENDAR
VERSION:2.0
X-WR-CALNAME:EventsCalendar
PRODID:-//hacksw/handcal//NONSGML v1.0//EN
CALSCALE:GREGORIAN
BEGIN:VTIMEZONE
TZID:America/New_York
LAST-MODIFIED:20240422T053451Z
TZURL:https://www.tzurl.org/zoneinfo-outlook/America/New_York
X-LIC-LOCATION:America/New_York
BEGIN:DAYLIGHT
TZNAME:EDT
TZOFFSETFROM:-0500
TZOFFSETTO:-0400
DTSTART:19700308T020000
RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU
END:DAYLIGHT
BEGIN:STANDARD
TZNAME:EST
TZOFFSETFROM:-0400
TZOFFSETTO:-0500
DTSTART:19701101T020000
RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
CATEGORIES:College of Engineering,Thesis/Dissertations
DESCRIPTION:College of Engineering Computer & Information Science Location:
  https://umassd.zoom.us/j/99756832558?pwd=5Jj2KWIfqIAz7H1P7QiC8E9LgRRWMm.1
  Thesis Advisor:  Dr. Ashokkumar Patel, CIS/DSC - University of Massachus
 etts, Dartmouth Committee Members:  Dr. Gokhan Kul, CIS/DSC - University o
 f Massachusetts, Dartmouth Dr. Yuchou Chang, CIS/DSC - University of Massa
 chusetts, Dartmouth  Abstract Credential misuse is hard to detect because 
 every login in an attack can succeed. This thesis asks whether judging eac
 h login against accounts with similar roles, its peer group, reduces false
  alarms at the same detection rate. A peer-aware detector is proposed: acc
 ounts are clustered on behavior and host usage, each group is modeled by i
 ts own isolation forest, and scores become peer-relative tail probabilitie
 s. It is compared with a first-time-host rule and with per-user, network-a
 ware and peer-column models on the Los Alamos National Laboratory authenti
 cation data, using three chronological partitions, causal features, a test
  set locked until all choices are frozen, and automated leakage checks. On
  a test period of 44.2 million logins with 333 red-team logins, peer featu
 res used as extra inputs raised average precision 1.7-fold over the networ
 k-aware model on every seed, and the peer-aware detector needed 13% fewer 
 alerts at exactly 80% recall, a difference not significant across seeds. N
 either advantage held at 90% recall, and calibration did not improve ranki
 ng. Thresholds set to catch 80% of validation attacks caught only 59–69%
  of test attacks, and grouping alerts per account, destination and day cut
  workload about fivefold. Peer context helps in specific places, but thres
 hold transfer between attack periods matters more for deployment than the 
 differences between detectors. For further information please contact Dr.
  Ashokkumar Patel.  All CIS MS students are encouraged to attend.  \nEve
 nt page: https://www.umassd.edu/events/cms/12-14-26-peer-group-aware-anoma
 ly-detection-for-credential-misuse.php\nEvent link: https://umassd.zoom.us
 /j/99756832558?pwd=5Jj2KWIfqIAz7H1P7QiC8E9LgRRWMm.1
X-ALT-DESC;FMTTYPE=text/html:<html><body><p>College of Engineering</p>\n<p>
 Computer & Information Science</p>\n<p>Location: <a href="https://umassd.z
 oom.us/j/99756832558?pwd=5Jj2KWIfqIAz7H1P7QiC8E9LgRRWMm.1">https://umassd.
 zoom.us/j/99756832558?pwd=5Jj2KWIfqIAz7H1P7QiC8E9LgRRWMm.1</a></p>\n<p>The
 sis Advisor:  Dr. Ashokkumar Patel\, CIS/DSC - University of Massachusett
 s\, Dartmouth</p>\n<p>Committee Members:</p>\n<ul>\n<li>Dr. Gokhan Kul\, C
 IS/DSC - University of Massachusetts\, Dartmouth</li>\n<li>Dr. Yuchou Chan
 g\, CIS/DSC - University of Massachusetts\, Dartmouth</li>\n</ul>\n<p>Abst
 ract</p>\n<p>Credential misuse is hard to detect because every login in an
  attack can succeed. This thesis asks whether judging each login against a
 ccounts with similar roles\, its peer group\, reduces false alarms at the 
 same detection rate. A peer-aware detector is proposed: accounts are clust
 ered on behavior and host usage\, each group is modeled by its own isolati
 on forest\, and scores become peer-relative tail probabilities. It is comp
 ared with a first-time-host rule and with per-user\, network-aware and pee
 r-column models on the Los Alamos National Laboratory authentication data\
 , using three chronological partitions\, causal features\, a test set lock
 ed until all choices are frozen\, and automated leakage checks. On a test 
 period of 44.2 million logins with 333 red-team logins\, peer features use
 d as extra inputs raised average precision 1.7-fold over the network-aware
  model on every seed\, and the peer-aware detector needed 13% fewer alerts
  at exactly 80% recall\, a difference not significant across seeds. Neithe
 r advantage held at 90% recall\, and calibration did not improve ranking. 
 Thresholds set to catch 80% of validation attacks caught only 59–69% of 
 test attacks\, and grouping alerts per account\, destination and day cut w
 orkload about fivefold. Peer context helps in specific places\, but thresh
 old transfer between attack periods matters more for deployment than the d
 ifferences between detectors.<br> <br>For further information please cont
 act Dr. Ashokkumar Patel.  All CIS MS students are encouraged to attend. 
  </p><p>Event page: <a href="https://www.umassd.edu/events/cms/12-14-26-p
 eer-group-aware-anomaly-detection-for-credential-misuse.php">https://www.u
 massd.edu/events/cms/12-14-26-peer-group-aware-anomaly-detection-for-crede
 ntial-misuse.php</a><br>Event link: <a href="https://umassd.zoom.us/j/9975
 6832558?pwd=5Jj2KWIfqIAz7H1P7QiC8E9LgRRWMm.1">https://umassd.zoom.us/j/997
 56832558?pwd=5Jj2KWIfqIAz7H1P7QiC8E9LgRRWMm.1</a></p></body></html>
DTSTAMP:20261119T144038
DTSTART;TZID=America/New_York:20261214T130000
DTEND;TZID=America/New_York:20261214T140000
LOCATION:Online - Zoom
SUMMARY;LANGUAGE=en-us:Peer-Group-Aware Anomaly Detection for Credential Mi
 suse in Enterprise Authentication Logs
UID:60326c29f6211751c2c6fc5d7da77232@www.umassd.edu
END:VEVENT
END:VCALENDAR
