Sign inSign up
Deep-er Kernels thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Deep-er Kernels

Published on Aug 26, 201312001 Views

John Shawe-Taylor

Kernels can be viewed as shallow in that learning is only applied in a single (output) layer. Recent successes with deeper networks highlight the need to consider richer function classes. The talk w

ROKS 2013 - Leuven

Related categories

Compressed SensingSupport Vector MachinesRegularizationOptimization Methods

Chapter list

Deep-er Kernels00:00
Background - 100:03
Background - 200:36
Background - 300:52
Background - 401:09
Why Shallow Learning? - 101:40
Why Shallow Learning? - 202:04
Why Shallow Learning? - 302:50
What happens in practice? - 103:48
What happens in practice? - 204:35
What happens in practice? - 306:07
Aim of this talk - 107:40
Aim of this talk - 207:53
Aim of this talk - 307:59
Aim of this talk - 408:11
Aim of this talk - 508:33
Aim of this talk - 608:43
Matching pursuit - 108:52
Matching pursuit - 209:44
Matching pursuit bound plot12:23
Matching pursuit - 313:07
Kernels from Probabilistic Models - 113:27
Kernels from Probabilistic Models - 213:55
Kernels from Probabilistic Models - 314:59
Fisher kernels - 116:02
Fisher kernels - 216:45
Fisher kernels - 316:58
Fisher kernels - 417:20
Fisher kernels - 517:44
String kernels as Fisher kernels - 119:25
String kernels as Fisher kernels - 220:07
String kernels as Fisher kernels - 322:01
String kernels as Fisher kernels - 422:38
Multiple kernel learning - 125:50
Multiple kernel learning - 227:50
Rademacher complexity28:11
Bounding MKL - 130:58
Bounding MKL - 231:24
Bounding MKL - 331:55
Bounding MKL - 432:48
Bounding MKL - 533:16
Experimental results with large-scale MKL34:20
Linear programming boosting35:47
MKL Algorithmics - 139:41
Linear Programming MKL39:50
MKL Algorithmics - 240:02
MKL Algorithmics - 340:08
MKL Algorithmics - 440:32
Learning Fisher kernels - 141:11
Learning Fisher kernels - 241:22
Non-linear Feature Selection - 141:56
Non-linear Feature Selection - 243:04
Analysis - 145:24
Analysis - 245:30
Analysis - 345:33
Analysis - 445:48
Analysis - 545:52
Example46:01
Results - 146:11
Results - 246:55
Results - 347:00
Results - 447:09
Results - 547:15
Results - 647:19
Results - 747:25
Summary and Conclusions - 147:42
Summary and Conclusions - 247:47
Summary and Conclusions - 347:53
Summary and Conclusions - 447:58