Multimodal Learning with Deep Boltzmann Machines thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Multimodal Learning with Deep Boltzmann Machines

Published on Jan 16, 201314976 Views

We propose a Deep Boltzmann Machine for learning a generative model of multimodal data. We show how to use the model to extract a meaningful representation of multimodal data. We find that the lear

Related categories

Chapter list

Mul$modal Learning with Deep Boltzmann Machines00:00
Data – Collection of Modalities (1)00:24
Data – Collection of Modalities (2)00:46
Shared Concept00:57
Building a Probabilistic Model (1)01:16
Building a Probabilistic Model (2)01:36
Building a Probabilistic Model (3)01:43
Challenges ­‐ I02:01
Challenges -­ IIa02:39
Challenges -­ IIb03:07
Genera$ng Text from Images03:24
Restricted Boltzmann Machines04:27
RBMs for Real-­valued Data05:12
RBMs for Word Counts05:49
A Nice Thing about RBMs06:22
A Simple Multimodal Model06:41
Deep Boltzmann Machines06:50
Multimodal DBM (1)07:14
Multimodal DBM (2)07:42
Multimodal DBM (3)08:03
Multimodal DBM (4)08:25
Learning DBMs (1)08:45
Learning DBMs (2)09:03
Text Generated from Images (1)09:51
Text Generated from Images (2)10:15
Images from Text (1)11:00
Images from Text (2)11:49
Pretraining12:12
MIR‐Flickr Dataset12:30
Data and Architecture12:46
Results (1)13:21
Results (2)14:01
Results (3)14:14
Results (4)14:28
Thank you!15:08