BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Memento EPFL//
BEGIN:VEVENT
SUMMARY:Multi-Task Learning: a Bayesian approach
DTSTART:20121218T130000
DTEND:20121218T140000
DTSTAMP:20260407T143843Z
UID:d2d5ef96f2ab6f8838b2085b5f5b9f465d6de07eabfdc37028596fcd
CATEGORIES:Conferences - Seminars
DESCRIPTION:Dr. Cedric Archambeau\,  Machine Learning for Services group 
 at Xerox Research Center Europe\nIn many real life problems multiple relat
 ed target variables need to be predicted from a single set of input featur
 es. A problem that attracted considerable interest in recent years is to l
 abel an image with (text) keywords based on the features extracted from th
 at image. In general\, this multi-label classification problem is challeng
 ing as the number of classes is equal to the vocabulary size and thus typi
 cally very large. While capturing correlations between the labels seems ap
 pealing it is in practice difficult as it rapidly leads to numerical probl
 ems when estimating the correlations. In this talk\, I will introduce a sp
 arse Bayesian model for multi-task regression and classification. The mode
 l is able to capture correlations between tasks\, while being sparse in th
 e features. This is especially attractive for the interpretation of the re
 sults. By adopting a Bayesian approach we can learn the level of sparsity 
 from the data. Empirical evaluations on data sets from biology and vision 
 demonstrate the applicability of the model\, where on both regression and 
 classification tasks it achieves competitive predictive performance compar
 ed to previously proposed methods. I will also briefly discuss how the mod
 el was applied to the analysis of data collected in Xerox call centres.\nT
 his is joint work with Shengbo Guo and Onno Zoeter.
LOCATION:INR113 http://plan.epfl.ch/?zoom=20&recenter_y=5863814.94355&rece
 nter_x=730548.85489&layerNodes=fonds\,batiments\,labels\,events_surface\,e
 vents_line\,events_label\,information\,parkings_publics\,arrets_metro\,eve
 nem
STATUS:CONFIRMED
END:VEVENT
END:VCALENDAR