BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:IC Colloquium : Fast learning algorithms for discovering the hidde
 n structure in data
DTSTART:20130207T161500
DTEND:20130207T173000
DTSTAMP:20260407T050809Z
UID:9acd81f8febe096a8239ac6b0ac86bfa46a0b55d1387c92c4bfb561d
CATEGORIES:Conferences - Seminars
DESCRIPTION:Daniel Hsu\, Microsoft Research\nIC faculty candidate\nAbstrac
 t\nA major challenge in machine learning is to reliably and automatically 
 discover hidden structure in data with little or no human intervention. Ma
 ny of the core statistical estimation problems of this type are\, in gener
 al\, provably intractable for both computational and information-theoretic
  reasons.  However\, much progress has been made over the past decade or 
 so to overcome these hardness barriers by focusing on realistic cases that
  rule out the intractable instances.  In this talk\, I'll describe a gene
 ral computational approach for correctly estimating a wide class of statis
 tical models\, including Gaussian mixture models\, Hidden Markov models\, 
 Latent Dirichlet Allocation\, Probabilistic Context Free Grammars\, and se
 veral more. The scope of the approach extends beyond the purview of previo
 us algorithms\; and it leads to both new theoretical guarantees for unsupe
 rvised learning\, as well as fast and practical algorithms for large-scale
  data analysis.Biography\nDaniel Hsu is a postdoc at Microsoft Research Ne
 w England.  Previously\, he was a postdoc with the Department of Statisti
 cs at Rutgers University and the Department of Statistics at the Universit
 y of Pennsylvania from 2010 to 2011\, supervised by Tong Zhang and Sham M.
  Kakade.  He received his Ph.D. in Computer Science in 2010 from UC San D
 iego\, where he was advised by Sanjoy Dasgupta\; and his B.S. in Computer 
 Science and Engineering in 2004 from UC Berkeley.  His research interests
  are in algorithmic statistics and machine learning.
LOCATION:BC 420 https://plan.epfl.ch/?room==BC%20420
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR