LolWolf

50 ms·

241.
by LolWolf 7y ago
While this is a nice visualization of the SIR model (which is what I think this is), the lack of variation in transmission rates (and no small-world behavior, which is what we'd expect in many cases relating to the current world), make
242.
by LolWolf 7y ago
I first guessed it was the chemical, then remembered I was on HN, and did a double take also thinking it was a Rust rewrite of some fancy library, or something. I was (pleasantly?) surprised at the outcome, I guess...
243.
by LolWolf 7y ago
Still, the test specificity (or more generally, its predictive power) is likely to be quite low simply due to the sheer number of flu cases. Otoh, this doesn’t really explain why there would be regulations against developing in-lab tests (a
244.
by LolWolf 7y ago
Heya! Very nice and congratulations on the front page :) On (2) that's the great part about solvers, is that they're essentially quite incredible black boxes. Most people who actually do optimization theory also don't really
245.
by LolWolf 7y ago
Oh, oops, I just realized I actually wrote "comic," oops! (My phone really does not like the word conic for whatever reason...) Would be a fun (and uhh, interesting?) challenge indeed, whatever that would entail :)
246.
by LolWolf 7y ago
I did mean to write mixed integer convex optimization; on the other hand, conic programming includes convex programming as a special case (and vice versa in the case where cones are convex). Most convex problems we solve (e.g. When written
247.
by LolWolf 7y ago
There are also ways around this by suitable generalizations of MILPs, namely mixed integer comic programs (MICPs), which are about as fast to solve in practice. This generalization allows you to use nonlinear (convex) functions in the objec
248.
by LolWolf 7y ago
Sure, the SAT -> ILP is small but the reduction to SAT—or MILP directly—can sometimes (but not often) be large. Either way, I agree that there is likely some way in which the original commenter’s problem can be solved via MILP/MICP
249.
by LolWolf 7y ago
Yeah, but the problem is that the reduction might be large in practice (in the sense that reducing the original problem to SAT or MIP might require the introduction of a large number of variables). While theoretically polynomially reducible
250.
by LolWolf 7y ago
Depends on the problem (in max-flow, min-cut the trivial rounding scheme gets you the optimal point immediately :). Kidding aside, generally this is very true, but for a surprising number of practical problems, the LP relaxation and some re
251.
by LolWolf 7y ago
I’d be curious to hear more... often a lot of nonlinear problems can be turned into mixed integer cone programs (MICPs) which are also often fast to approximately or exactly solve in practice.
252.
by LolWolf 7y ago
This is a lovely example, thank you for the link!
253.
by LolWolf 7y ago
The shape of the bokeh is given by the shape of the aperture, so it is also a property of the lenses. The specific word is related to the aesthetic quality of the effect in context, but that doesn’t mean it is truly independent of the lens.
254.
by LolWolf 7y ago
Still doesn't fix the inverse mapping problem, which is done immediately by Autoencoders, by definition.
255.
by LolWolf 7y ago
UMAP is, sadly, extraordinarily slow, plus the inverse mapping is not clear.
256.
by LolWolf 7y ago
You should check out TexPad![0] It’s lovely and essentially real-time, with autocomplete, etc. I use it as my editor of choice for all of my papers (not that any of them have particularly challenging layouts). Really cannot recommend it eno
257.
by LolWolf 7y ago
Hmm, in most embedded processors I know (e.g., the ESP8266/32 [0], RPi's BMC2835 [1], or Ti's MSP430 [2]), you send the line down for reset but hold it high for running (either by internal pull-up or by actually holding the
258.
A mathematical analysis of Uniswap markets(arxiv.org)
30 points by LolWolf 7y ago|6 comments
259.
by LolWolf 7y ago
Oh sure, I agree with that. For more context, I highly recommend chapter 5 in B&V.[0] —— [0] http://web.stanford.edu/~boyd/cvxbook/bv_cvxbook.pdf#page229
260.
by LolWolf 7y ago
Turning a problem into a dual version is mostly a mechanical thing—you can immediately write it down in the case of LPs, for example[0], and many other convex cases [1]. —— [0] https://en.m.wikipedia.org/wiki/Dual_linea
261.
by LolWolf 7y ago
I mean, perhaps; though I've found it does quite the opposite. The idea is proposed front and center, sure (this approach does not preclude that), but far less time is spent explaining the motivation and doing mathematical yoga to clar
262.
by LolWolf 7y ago
I think the GP is trying to say that, once you get past all of the definitions and proofs, the underlying idea is almost always relatively simple. I do agree that it's pretty hard to be rigorous without setting up all of this edifice,
263.
by LolWolf 7y ago
To be fair, this is one point that I think they've actually gotten correct. Discrete log is easy <=> Integer factorization is easy.
264.
The quest to beat electronics with photonics(osa-opn.org)
2 points by LolWolf 7y ago|0 comments
265.
by LolWolf 7y ago
The algorithm is a randomized algorithm and we're assuming we're picking a good hash. The probability of that event is exponentially vanishing, so, in expectation (which is the runtime we're looking for, as we are analyzing a
266.
by LolWolf 7y ago
I'm still not sure about this—the amortized time of insertion should be O(1) on any one of the usual hash table-type-structures. We're not inserting a single item to an already-built hash-table, but rather k of them into an empty
267.
by LolWolf 7y ago
Indeed, but this assumes that modern crypto is secure! :)
268.
by LolWolf 7y ago
In general, assuming that we have constant-time operations on arbitrarily large numbers (which we almost universally assume when doing these analyses, except in very specific fields), then hash structures do check membership and add items i
269.
by LolWolf 7y ago
It's... not that useful, as far as I can tell (unless I'm still under-caffeinated!) @avip mentions the following comment https://news.ycombinator.com/item?id=20961320 , which means that we only have to consider k
270.
by LolWolf 7y ago
See my edit! Oops. Anyways, if you were to keep it in some hash-based structure, you could check membership! (But I'm sure you already knew that :)

More ›