BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Higher Order Learning for Classification in Emergency Situations -
  Hannah Pauline Keiler (Columbia University and DIMACS)
DTSTART:20130806T100000Z
DTEND:20130806T110000Z
UID:TALK46514@talks.cam.ac.uk
CONTACT:Zoubin Ghahramani
DESCRIPTION:Traditional machine learning methods (like Naïve Bayes and La
 tent Dirichlet Allocation) only consider relationships between feature val
 ues within individual data instances while disregarding the dependencies t
 hat link features across instances. My group at DIMACs has developed Highe
 r Order Learning -- a general approach to supervised learning that leverag
 es higher-order dependencies between features across instances. Higher Ord
 er Learning has been shown to outperform traditional machine learning tech
 niques in multiple accounts. This talk will focus on two Higher Order Lear
 ning methods: Higher Order Naïve Bayes (HONB) and Higher Order Latent Dir
 ichlet Allocation (HOLDA). \nMore specifically\, my group has been interes
 ted in applications of HONB and HOLDA to situations concerning national se
 curity and emergency response. Some examples of this include nuclear detec
 tion and classification of needs during an emergency using crowdsourced da
 ta from social media and text messages. \n
LOCATION:Engineering Department\, CBL Room BE-438
END:VEVENT
END:VCALENDAR