[Colloquium] TTI-C Learning Theory Program Talk 3/9/ @ 3:00pm

[Katherine Cumming]

TOYOTA TECHNOLOGICAL INSTITUTE 
Learning Theory Program
 
Wednesday, March 9th, 3:00 pm
TTI-C Conference Room (1427 E. 60th St. - 2nd Floor) 
 
 
 
Speaker: Lorenzo Rosasco, University of Genoa
Speaker's homepage: http://www.disi.unige.it/person/RosascoL/
 
Title:  From Interpolation to Learning from Examples
Abstract:
In a deterministic setting the problem of interpolation and
approximation of a multivariate function can be easily cast in the
framework of ill-posed inverse problems. In this context the notion of
RKHS naturally arises and regularization techniques allow to find
meaningful solution to the problem. 

It is well known that learning from examples can be seen as a
multivariate function approximation problem. Nonetheless the
probabilistic nature of the learning does not allow a straightforward
mathematical connection with inverse problems theory. We present a
possible way to tackle this problem for the specific case of square
loss. Formal connections between the consistency approach imposed in
learning theory, and the regularization-stability convergence property
used in ill-posed inverse problems can be derived. 
If you have questions, or would like to meet the speaker, please contact
Katherine at 4-1994 or kcumming at tti-c.org. For information on future
TTI-C talks or events, please go to the TTI-C Events
<http://ttic.uchicago.edu/events/events_dyn.php>  page. 
 
 
 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.cs.uchicago.edu/pipermail/colloquium/attachments/20050307/b5962cb0/attachment.htm