Best Paper - Information-Theoretic Metric Learning thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Best Paper - Information-Theoretic Metric Learning

Published on Jun 22, 200717774 Views

In this paper, we present an information-theoretic approach to learning a Mahalanobis distance function. We formulate the problem as that of minimizing the differential relative entropy between two mu

Related categories

Chapter list

Information-theoretic Metric Learning00:00
Introduction00:21
Our approach01:37
Mahalanobis distances02:37
Problem formulation04:06
The Gaussian connection05:36
The optimization problem-part0106:55
The optimization problem-part0208:01
Bergman's method-part0109:02
Bergman's method-part0209:48
Connection to Kernel learning10:18
Kernelization12:01
Online metric learning13:11
Experimental results14:57
UCI data sets15:47
Clarify data sets16:29
Conclusions17:45