BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:IC Mondays seminars - Algorithmic Challenges in Machine Learning
DTSTART:20100322T161500
DTSTAMP:20260407T050811Z
UID:8197cf84de5b111950d240756820313386cc9228635c8225f7e27be8
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr. Kamalika Chaudhuri\, UCSD\, Computer Science Department\nA
 bstract\n\nIn this talk\, we address two algorithmic challenges in machine
  learning.\n\nFirst\, with the increase in electronic record-keeping\, man
 y datasets that learning algorithms work with relate to sensitive informat
 ion about individuals. Thus the problem of privacy-preserving learning -- 
 how to design learning algorithms that operate on the sensitive data of in
 dividuals while still guaranteeing the privacy of individuals in the train
 ing set -- has achieved great practical importance. In this talk\, we addr
 ess the problem of privacy-preserving classification\, and we present an e
 fficient classifier which is private in the differential privacy model of 
 Dwork et al. Our classifier works in the ERM (empirical loss minimization)
  framework\, and includes privacy preserving logistic regression and priva
 cy preserving support vector machines. We show that our classifier is priv
 ate\, provide analytical bounds on the sample requirement of our classifie
 r\, and evaluate it on some real data.\n\nSecond\, we address the problem 
 of clustering\, when data is available from multiple domains or views. For
  example\, when clustering videos by speaker\, we have access to the audio
  stream as well as the images.\nIn this talk\, we address this problem of 
 Multiview Clustering\, and show how to use information from multiple views
  to improve clustering performance. We present an algorithm for multiview 
 clustering\, provide analytical bounds on the performance of our algorithm
  under certain statistical assumptions\, and finally evaluate our algorith
 m on some real data.\n\nBased on joint work with Sham Kakade (UPenn)\, Kar
 en Livescu (TTI Chicago)\, Claire Monteleoni (CCLS Columbia)\, Anand Sarwa
 te (ITA UCSD)\, and Karthik Sridharan (TTI Chicago).\n\nBiography\n\nKamal
 ika Chaudhuri received a Bachelor of Technology degree in Computer Science
  and Engineering in 2002 from the Indian Institute of Technology\, Kanpur\
 , and a PhD in Computer Science from UC Berkeley in 2007. She is currently
  a postdoctoral researcher at the Computer Science and Engineering Departm
 ent at UC San Diego.\n\nKamalika's research is on the design and analysis 
 of machine-learning algorithms and their applications. In particular\, her
  interests lie in  clustering\, online learning\, and privacy-preserving m
 achine-learning\, and applications of machine-learning and algorithms to p
 ractical problems in other areas.
LOCATION:INM 202
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR