[Colloquium] Daniel Hsu will visit Wednesday, February 27, 2013. His talk will be about "Fast Learning algorithms for discovering the hidden structure in data" Please join us in Ryerson 251 at 2:30 pm

[Natalie Cruz]

The University of Chicago Computer Science Department

 

PRESENTS:

 

“Fast learning algorithms for discovering the hidden structure in data”

 

 

Daniel Hsu

Postdoctoral Researcher

Microsoft Research

Cambridge University

 

Abstract:

 

	A major challenge in machine learning is to reliably and automatically discover hidden structure in data with little or no human intervention. Many of the core statistical estimation problems of this type are, in general, 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 general computational approach for correctly estimating a wide class of statistical models, including Gaussian mixture models, Hidden Markov models, Latent Dirichlet Allocation, Probabilistic Context Free Grammars, and several more.  The key idea is to exploit the structure of low-order correlations that is present in high-dimensional data.  The scope of the new approach extends beyond the purview of previous algorithms; and it leads to both new theoretical guarantees for unsupervised learning, as well as fast and practical algorithms for large-scale data analysis.


 

Wednesday, February 27th, 2013

2:30 PM

Ryerson 251

Host: John Lafferty

 

People in need of assistance should call Natalie Cruz (773) 834-8977 in advance.

For additional information on future CS talks please visit: http://www.cs.uchicago.edu/events
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20130225/2183af9f/attachment-0001.htm 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Daniel Hsu.png
Type: image/png
Size: 56167 bytes
Desc: not available
Url : http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20130225/2183af9f/attachment-0001.png