Since the video likes to show slides for a few seconds here are some timestamps. I only included the first time a slide showed up Goals: 2:10 Topic: Dual decomposition Dual problem: 6:21 Dual ascent: 8:25 Dual decomposition: 10:28, 14:09 Method of Multipliers: 15:05 Comparing Method of Multipliers to Dual Decomposition:19:19 Alternating Direction Method of Multipliers: 24:39 Convergence: 30:50 ADMM with scaled dual variables: 37:07 Related algorithms: 39:09 Common Patterns: Decomposition: 42:00 Proximal operator: 43:59 Quadratic objective: 44:10 Smooth objective: 49:08 Examples: Constrained convex optimization: 50:35 Lasso: 55:49 Sparse inverse covariance selection: 59:48 Sparse inverse covariance selection via ADMM: 1:00:46 Consensus and exchange: Consensus Optimization: 1:02:29 Consensus SVM example: 1:13:59 Distributed LASSO example: 1:16:48 Consensus Optimization via ADMM: 1:18:28 Exchange ADMM: 1:22:48 Exchange problem: 1:23:19 Solve time scaling: 1:28:07 Conclusions: Summary and conclusions: 1:31:01
It would be helpful if the slides were shown more. I had to pause every time the slides were shown for only a few seconds to get what he was talking about.
Thanks for the talk. It's quite clear and useful. It would be great if you uploaded the slides, it's not practical to perceive a slide for a couple of seconds only.
quite interesting, but we're not here to admire the speaker's shirt ... showing the entire room with the slides (and so the laser pointer) would have more meaning ...
Since the video likes to show slides for a few seconds here are some timestamps. I only included the first time a slide showed up
Goals: 2:10
Topic:
Dual decomposition
Dual problem: 6:21
Dual ascent: 8:25
Dual decomposition: 10:28, 14:09
Method of Multipliers: 15:05
Comparing Method of Multipliers to Dual Decomposition:19:19
Alternating Direction Method of Multipliers: 24:39
Convergence: 30:50
ADMM with scaled dual variables: 37:07
Related algorithms: 39:09
Common Patterns:
Decomposition: 42:00
Proximal operator: 43:59
Quadratic objective: 44:10
Smooth objective: 49:08
Examples:
Constrained convex optimization: 50:35
Lasso: 55:49
Sparse inverse covariance selection: 59:48
Sparse inverse covariance selection via ADMM: 1:00:46
Consensus and exchange:
Consensus Optimization: 1:02:29
Consensus SVM example: 1:13:59
Distributed LASSO example: 1:16:48
Consensus Optimization via ADMM: 1:18:28
Exchange ADMM: 1:22:48
Exchange problem: 1:23:19
Solve time scaling: 1:28:07
Conclusions:
Summary and conclusions: 1:31:01
Much helpful.. Thank you !
Slides: web.stanford.edu/class/ee364b/lectures/admm_slides.pdf
It would be helpful if the slides were shown more. I had to pause every time the slides were shown for only a few seconds to get what he was talking about.
Ahmad Bazzis lectures are much more clear and concise.
Could you please send the link of Ahmad Bazzis lecture on this topic (ADMM)?
I'm sorry to dislike such video...but...i don't care to see boyd's face. I want to see the slides
Yeah what kind of a MANIAC edited this video
Thanks for the talk. It's quite clear and useful. It would be great if you uploaded the slides, it's not practical to perceive a slide for a couple of seconds only.
Thank you :)
Can you upload the video from the camera that shows the lecturer and the slides?
Sorry that the recording did not follow the presentation!
What is the definition of "CG step" (discussed on 47:55 )?
is it conjugate gradient? not sure
(Probably) the slides in the talk: web.stanford.edu/~boyd/papers/pdf/springer_15_lect2.pdf
good explanation. but the camera is only focus on the advisor only. bad thing
can't see the slides -.-
Great Lecture
One question.
What is the code of the "pos()" in matlab from your ADMM examples
z = pos(x_hat + u);
Basicly what the pos() does ?
pos(x) = x_+ = max{0, x}
i like this guy, any more vids with him?
there are 40 videos of his lectures here on yt. look for "convex optimization"
isn't like an insualt to on
Microsoft and use a mac laptop?
it works if you get the slide number from the right-hand footer next to boyd's head
i wish, someone had told the cameraman to focus on slides when he would explain.
It must be a business people doing the edit.
Is camera sleeping?
quite interesting, but we're not here to admire the speaker's shirt ... showing the entire room with the slides (and so the laser pointer) would have more meaning ...
360p, really?
Camare should be on the slides
Great talk but the way the video is presented is simply terrible. The slides need to be shown along with the presenter.
It would be great if the captains are available.
where is the slides?
Video making is not good. Only focused to speaker instead of slides.
This video is almost useless. It shows Stephen Boyd for 95 percent of the time and the formulas 5 percent of the time...
The cameraman is very poor.
This is not useful without the slides
I dislike this video only because it did not show the slides!