0.25
0.5
0.75
1.25
1.5
1.75
2
Conformal Multi-Instance Kernels
Published on Feb 25, 20074910 Views
In the multiple instance learning setting, each observation is a bag of feature vectors of which one or more vectors indicates membership in a class. The primary task is to identify if any vectors
Related categories
Chapter list
Conformal Multi-Instance Kernels00:00
Outline00:12
Multiple Instance Learning00:53
Related Work02:00
Related Work (continued)03:07
Kernels between distributions04:25
Kernel Density Estimation Over Bags06:44
Conformal Kernels (Amari and Wu, 1999)07:17
Conformal Multi-Instance Kernels09:01
Implementation details09:41
Gradient Descent on the Radius-Margin Bound11:15
Optimizing the Trace-Margin Bound11:43
Diagnolization of conformal transformation12:28
A toy example13:40
Experimental Results16:02
Future Work18:20
Thank you20:01
Diagnolization of conformal transformation0122:07