LolWolf

19 ms·

181.
by LolWolf 6y ago
Yes! Well, at least, for the most part. I am also coincidentally paid whenever I discover math and explain it to an audience, but this case is much rarer than the former.
182.
by LolWolf 6y ago
Hmm... I'm not sure I agree. All optimal points (for, say, optimizing a linear function) will lie on the extremal points of the feasible domain, many of which will be points where the constraint functions are not differentiable. In all
183.
by LolWolf 6y ago
Not sure if it counts as "free time" as I study math for a living :) but optimization theory is always on my list. If you like Strang's new book, I think you'll be quite partial to Boyd's VMLS [0] which is (in my ad
184.
by LolWolf 6y ago
Depending on the level you're looking for, I'm always a huge fan of Rockafellar [0] for the more mathematical expositions (they're lovely, clear, and very well thought through). Even in this case, though, subgradients may not
185.
by LolWolf 6y ago
Well, every closed, Markovian system is governed by thermodynamics in a rather specific mathematical sense [0]; but yes I agree that initial conditions and parameters (such as constants) are not, but the dynamics themselves are. Unless our
186.
by LolWolf 6y ago
Hi! One of the authors here. We thought this paper might be of interest to people in data analysis or machine learning for fitting problems which have many outliers in data (in relatively simple ways). We a few examples where most common he
187.
Minimizing a sum of clipped convex functions(arxiv.org)
3 points by LolWolf 6y ago|1 comments
188.
by LolWolf 6y ago
Hey, Kerbal Space Program is a perfectly valid alternative to a degree in Aerospace engineering! EDIT: Obviously kidding.
189.
by LolWolf 6y ago
Then you boot up any of the hyper-realistic WWII ARMA mods and you’re set.
190.
by LolWolf 6y ago
Very. I read the article and was very confused for quite a while until I realized this.
191.
by LolWolf 6y ago
Oh it’s an absolutely lovely overview. Honestly I would highly recommend most stuff from Diaconis, he’s got a fantastic expository style (and all of his papers are quite readable!)
192.
by LolWolf 6y ago
Indeed, the closest things I've been able to find have been these two studies [0, 1] which investigate a large number of different possible incidental stimuli (from lights all the way to the placement of the assistants during the exper
193.
by LolWolf 6y ago
Awesome, thank you!
194.
by LolWolf 6y ago
What do you mean by filter analysis? As in classical linear filtering? (as in, you're finding the roots of the transfer function?)
195.
by LolWolf 6y ago
I love it! I had heard of some of these results in relation to Persi Diaconis's work ( e.g. , [0]). I'm curious, what is the textbook? Would love to take a quick glance at it at least! :) EDIT: I also really like the fact that eve
196.
by LolWolf 6y ago
> that abuses appropeiated quantum physics terminology I would certainly not only specifically not say this is borrowed from quantum physics (perhaps classical EM?), but in fact the whole point of the theorem is to show that such roots
197.
by LolWolf 6y ago
Absolutely agreed—it's a little funny to see dismissals of the article (which, as I mentioned elsewhere in the thread is actually quite good), even though, as a mathematician [0] we use these analogies (and sometimes really the whole i
198.
by LolWolf 6y ago
Wow, this is an absolutely lovely presentation of that result. Huge props to Hartnett for writing this piece! It's a perfect mix of intuitive and well-explained without being too hand-wavy and it's quite an interesting subject, to
199.
by LolWolf 6y ago
You can absolutely do this (and it would be quite interesting to try with a VR headset, e.g.!); the idea is a special case of the projective transform [0] in 4 dimensions. It may make for a cool demo. ----- [0] https://en.wikip
200.
by LolWolf 6y ago
These are great, thanks!
201.
by LolWolf 6y ago
Oooh, I did not know this and love to read more about it! Do you have a good source/expository article?
202.
by LolWolf 6y ago
Well, I'm not sure what you mean by "concave part of a Pareto curve." (Perhaps you meant convex?) Usually, the Pareto curve is the epigraph of the individual function evaluations at the optimal points, so, unless your problem
203.
by LolWolf 6y ago
> The simple reason is that that illustration shows how we regularize models conceptually, with hard constraints, not how we actually implement regularization, with soft constraints! Note that these are equivalent. In particular, assumin
204.
by LolWolf 6y ago
Interesting. My dad is Uruguayan and I have not heard him use it (perhaps a biased sample?). I’m Venezuelan, for context, so it may explain this view :)
205.
by LolWolf 6y ago
That's pretty much spot on (more generally, I think it's a Central-American thing). Either "pelotas" or "bolas" are the common ones in South America (huevos is present, but rarely used in the same way—"hue
206.
by LolWolf 6y ago
Oh I know, but your shots specifically are lovely! I will check it out :) I imagined you did it at home... have wanted to do the same but haven’t had space, so use the mail-in services which are not bad.
207.
by LolWolf 6y ago
Oh my god that is a beautiful shot! The rest of these are also gorgeous... how do you develop them? / do you do prints at all?
208.
by LolWolf 6y ago
Gah, I've passed the editing window. If it's confusing to read, it is supposed to say: > the specific construction lets you pinpoint the people who come up positive
209.
by LolWolf 6y ago
This idea is called group testing [0] and has been used by the army for quite a while. It is also a well-studied problem in information theory as a (useful) basic model and we know optimal solutions to it. What is surprising is that group t
210.
by LolWolf 6y ago
> leads to the least certain outcome What's the loss function for this? How do you cross-validate this loss? The only generally applicable methods I can think of (essentially some Bayesian-type updating, which then uses Monte-Carlo-

More ›