en
0.25
0.5
0.75
1.25
1.5
1.75
2
Multiple-Instance Learning with Instance Selection via Dominant Sets
Published on Oct 17, 20113567 Views
Multiple-instance learning (MIL) deals with learning under ambiguity, in which patterns to be classified are described by bags of instances. There has been a growing interest in the design and use of
Related categories
Chapter list
Multiple-Instance Learning with Instance Selection via Dominant Sets00:00
Multiple-Instance Learning (MIL)00:14
Instance-Selection based MIL01:38
A comparison of Instance-Selection based MIL methods02:47
Dominant Sets - 106:49
Dominant Sets - 207:33
Dominant Sets - 308:32
MIL with Instance Selection via Dominant Sets (MILDS)09:31
MILDS - Basic Notations10:41
MILDS - Instance Selection (1)11:24
MILDS - Instance Selection (2)13:00
MILDS - Instance Selection (3)13:48
Classification15:04
One-vs-rest Multi-Class MILDS15:58
milDS (1)16:39
milDS (2)17:19
Experiments18:15
Benchmark Data Sets18:39
Benchmark Data Sets – The Dimensions of the Embedding Spaces19:41
Image Categorization - 120:58
Image Categorization - 222:00
Image Categorization – Selected Instance Prototypes (MILDS) (1)22:32
Image Categorization – Selected Instance Prototypes (MILDS) (2)23:20
Sensitivity to Labeling Noise24:42
Summary and Future Directions25:57
Any questions?27:23