Logo

Harvard Learning Seminar - Shared screen with speaker view
Depen Morwani cs19s002
22:34
Does g depend on inputs? if not then it can be taken into the weight matrix
Depen Morwani cs19s002
23:28
yes
Depen Morwani cs19s002
23:29
thanks
Depen Morwani cs19s002
39:07
But, the singular vectors of the matrices do change?
Depen Morwani cs19s002
39:18
so decoupled basis does change?
Depen Morwani cs19s002
39:56
thanks
Ankit Shukla
45:19
Was there a pattern observed whether certain 1D network gradient has a correlation with difficult of the examples ?
Ankit Shukla
47:09
Yep, meant difficult to classify
Ankit Shukla
48:06
Like images with blur or glare are difficult to classify
Boaz Barak
48:44
Maybe the examples we see now will help
Diganta Misra
59:21
Naive question - In a class incremental setting, when a new class arrives for a particular task - will it create a new pathway or just fine tune the existing pathway for that task?
Diganta Misra
01:00:51
Interesting, thanks :)
Diganta Misra
01:04:53
Follow up question - whether the learning dynamics on these pathway structures have been investigated under curriculum learning setting as compared to passing shuffled data samples during training?
Depen Morwani cs19s002
01:05:00
In the top plot, since the rates are comparable, why is XOR path completely suppressed?
Rainer Engelken
01:34:08
Great talk, can you comment on how/if this extends to recurrent networks?
Tariq Naeem
01:46:50
Great, talk!
Depen Morwani cs19s002
01:46:51
thanks a lot
Alessandro
01:46:54
Thank you
Depen Morwani cs19s002
01:46:55
:)
Tariq Naeem
01:46:59
Thank you.