Page killed my browser.
Why Momentum Works
71–80 of 100 posts
Re: Why Momentum Works
#72Earlier quoted context omitted.
Please check out our journal policies page: http://distill.pub/journal/ In brief, Distill needs to see three things to publish an article: outstanding communication, advancing the dialogue, and scientific integrity. Distill often works with authors to help them bring their articles up to our standards. Additionally, as a primary publication, Distill will not republish content already published elsewhere, or publish "…
Thank you! Of course I read it, just (as it is a new thing) still guessing what is a good fit, and what isn't. (For some reason I though that this t-SNE article was published elsewhere. Now I see that it was on Distill, but just before its big start.)
Yep, we're still clarifying our policies. If you have questions, please email us!
(Distill needs to be extra careful about a lot of this stuff because we're trying to build legitimacy for a kind of work that many people are inclined to not treat as academic contributions. So on some things, like being a primary publication, we may end up taking a more defensive posture than we would in an ideal world.)
Re: Why Momentum Works
#73Earlier quoted context omitted.
This is one of my favorite things ever. You should also look up "Jeremy England's Entropic Life" for a good companion thought.
A New Physics Theory of Life: https://www.quantamagazine.org/20140122-a-new-physics-theory... How Life (and Death) Spring From Disorder: https://www.wired.com/2017/02/life-death-spring-disorder/
Yep! I encountered both of these within a few weeks of each other, after having come to a similar (but way less educated and rigorous) conclusion to Prof England's.
Re: Why Momentum Works
#74Is the decreasing momentum related to the temperature in simulated annealing?
Re: Why Momentum Works
#75Earlier quoted context omitted.
Thank you! Of course I read it, just (as it is a new thing) still guessing what is a good fit, and what isn't. (For some reason I though that this t-SNE article was published elsewhere. Now I see that it was on Distill, but just before its big start.)
Yep, we're still clarifying our policies. If you have questions, please email us! (Distill needs to be extra careful about a lot of this stuff because we're trying to build legitimacy for a kind of work that many people are inclined to not treat as academic contributions. So on some things, like being a primary publication, we may end up taking a more defensive posture than we would in an ideal world.)
I wish all the best to this wonderful initiative! (As a side note, my co-advisor dreamt about this style of communication in science (though, in physics): https://physicsnapkins.wordpress.com/2012/04/16/a-personal-d...)
I totally understand that you need to be set high reference level at the very beginning, even if at the cost of being a bit "conservative" (I am not sure if the best word here).
Re: Why Momentum Works
#76Reminds me of this paper, "Causal Entropic Forces": http://math.mit.edu/~freer/papers/PhysRevLett_110-168702.pdf
This is one of my favorite things ever. You should also look up "Jeremy England's Entropic Life" for a good companion thought.
Relevant to this discussion about 'entropy and life' is the relation between the second law of thermodynamics and the purpose of life that this article presents : https://www.farnamstreetblog.com/2017/03/scientific-concepts...
Original article : https://www.edge.org/response-detail/27023
Re: Why Momentum Works
#77People, listen. "Damping" means to reduce the amplitude of an oscillation. "Dampening" means to make something wetter.
Re: Why Momentum Works
#78Pretty cool, it started using stuff about classical control theory. I always kinda missed that classical controls weren't really brought up in the discussion of gradient descent.
Re: Why Momentum Works
#79Earlier quoted context omitted.
The use of the term "accelerator" is misleading as it's introduced during the physical metaphor. You should consider editing that sentence to clarify your meaning: The added inertia smooths out variation in velocity, dampening oscillations and causing us to barrel through narrow valleys, small humps and local minima. Keeping our speed steadier, we arrive at the global optimum faster. Note the alliteration :-) > momen…
To the author: I found the sentences quoted above quite clear. Please do not change them. They helped me rapidly comprehend what the article was going to be about.
You don't find the idea that increased momentum causes increased speed a bit strange?
Re: Why Momentum Works
#80isn't this akin, in effect, to successive over relaxation? https://en.wikipedia.org/wiki/Successive_over-relaxation
under-relax, converge real slowly
over-over-relax, oscillate
just-right-over-relax, get fast convergence