Convergent Learning: Do different neural networks learn the same representations? thumbnail
slide-image
Pause
Mute
Subtitles not available
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Convergent Learning: Do different neural networks learn the same representations?

Published on May 27, 201610103 Views

Recent success in training deep neural networks have prompted active investigation into the features learned on their intermediate layers. Such research is difficult because it requires making sense o

Related categories

Chapter list

Convergent Learning: Do different neural networks learn the same representations? - 100:00
Convergent Learning: Do different neural networks learn the same representations? - 200:09
“Grandmother neuron”00:18
Local or Distributed? - 101:10
Local or Distributed? - 201:44
Second-last layer: Distributed02:15
Middle layers: sometimes local02:41
Local or Distributed? - 302:56
Local or Distributed? - 403:34
Local or Distributed? - 504:24
Concepts05:27
Key technique - 106:16
Key technique - 207:11
Key technique - 307:31
Key technique - 407:43
Neuron Activation Statistics: Correlation07:58
Aside: is correlation a powerful enough measure?08:43
One-to-one alignment between features - 109:11
One-to-one alignment between features - 209:49
One-to-one alignment between features - 310:04
One-to-one alignment between features - 410:36
One-to-one alignment between features - 510:42
One-to-one alignment between features - 610:48
One-to-one alignment between features - 711:07
One-to-one alignment between features - 811:09
One-to-one alignment between features - 911:16
One-to-one alignment between features - 1011:19
One-to-one alignment between features - 1111:22
One-to-one alignment between features - 1211:49
One-to-one alignment between features - 1311:55
One-to-one alignment between features - 1412:14
One-to-one alignment between features - 1512:18
One-to-one alignment between features - 1612:24
One-to-one alignment between features - 1712:41
One-to-one alignment between features - 1812:53
One-to-one alignment between features - 1913:16
One-to-one alignment between features - 2013:25
One-to-one alignment between features - 2113:33
One-to-one alignment between features - 2213:35
One-to-one alignment between features - 2313:53
One-to-one alignment between features - 2413:56
Untitled14:15
One-to-one alignment between features - 2614:17
One-to-one alignment between features - 2714:26
One-to-one alignment between features - 2814:31
One-to-one alignment between features - 2914:44
One-to-one alignment between features - 3014:54
One-to-one alignment between features - 3114:56
Few-to-one alignment between features - 115:01
Few-to-one alignment between features - 215:40
Few-to-one alignment between features - 316:26
Few-to-one alignment between features - 416:39
Few-to-one alignment between features - 516:48
Few-to-one alignment between features - 616:50
Few-to-one alignment between features - 716:52
Finding the low-dimensional subspaces - 117:05
Finding the low-dimensional subspaces - 217:11
Finding the low-dimensional subspaces - 317:23
Finding the low-dimensional subspaces - 417:25
Finding the low-dimensional subspaces - 517:34
Conclusions - 117:44
Conclusions - 218:02
Conclusions - 318:18
Conclusions - 418:38
Conclusions - 518:43
Untitled18:58
Contact - 118:59
Contact - 219:04
Thanks19:08