currymj

61 ms·

751.
by currymj 8y ago
Basically, it solves the problem of finding a setting of variables that is compatible with some set of constraints. This is NP-hard but Z3 is in practice very fast nonetheless. A case study/glowing review from a programmer at Microsoft
752.
by currymj 8y ago
this is extremely cool and well explained. the fact that the exploit comes literally in the form of shapes that can be drawn and visualized is very neat.
753.
by currymj 8y ago
Those are small examples but you can scale them up and Z3 can still handle it. I have no meaningful experience in OR, don’t know how SAT solvers work or how Z3 reduces things to SAT. Yet I was able to use it to attack a really gnarly schedu
754.
by currymj 8y ago
This is a kick ass book. It’s worth learning the basic use of some constraint solver, since they are now so fast and powerful. Z3 is free software and is very, very good, and you’d be surprised how many problems it can tackle (see: hundreds
755.
by currymj 8y ago
which sin was it?? greed would be the obvious, but one could make a case for pride and sloth as well.
756.
by currymj 8y ago
very stressful to try to teach under these conditions! the logistical problems of teaching and consistent grading grow worse than linearly with the number of students, unfortunately. if you want to have 500 people graded on the same curve,
757.
by currymj 8y ago
the elevator pitch is, it's an easy to use language, you can vectorize things if you want, but also if you hand-write loops it will be very fast. Whereas in R, hand-writing loops brings you to the 3rd circle of hell [1]. the more gener
758.
by currymj 8y ago
In the spirit of charity I wish there were a version of this post that didn’t seem like it was written after injecting some strange cocktail of anabolic steroids and cocaine. Maybe there are some good or at least interestingly wrong argumen
759.
by currymj 8y ago
idk if traditional security techniques can help with adversarial examples and that sort of thing. it is a very different kind of attack. training people to be paranoid and to worry about data provenance is important, though. i bet a bigger
760.
by currymj 8y ago
Video slot machines are particularly odious IMO. They are carefully designed to addict people. It's sort of like, don't ban tobacco, but maybe ban the worst tricks and chemicals cigarette companies use to make the product more add
761.
by currymj 8y ago
The first explicit reference I found was the Norwegian Scientific Index but I believe other countries have a similar system. Literally, each journal article is worth points, the points are divided among the coauthors, and then the points ha
762.
by currymj 8y ago
For the most part, researchers would prefer to publish open access, and agree with your stance about public money 100%, but with the current state of things, it would have a negative effect on their careers. You have to publish in certain h
763.
by currymj 8y ago
These were almost definitely not even Elsevier employees. They were editors of an Elsevier journal, but likely were paid only a token amount for their work, and probably not even that. The world of academic publishing is very strange, almos
764.
by currymj 8y ago
Yeah, I think the blog is written by lawyers who represent employers and/or draft noncompetes. Massachusetts recently almost banned them, California style, in response to the narrative that the Route 128 tech corridor could have been S
765.
by currymj 8y ago
the official party line is that the Cultural Revolution caused "the most severe setback and the heaviest losses suffered by the Party, the country, and the people since the founding of the People's Republic". many Communist p
766.
by currymj 8y ago
It seems like some people are actually pleased the hoax worked. If this is you, try to honestly imagine how you would feel if it had failed -- if the journals immediately called BS and refused to publish. Would you be happy or unhappy about
767.
by currymj 8y ago
It wasn't a conscious choice -- they were drafting their laws from a set of model statutes called the Field Code and simply didn't even think about changing the provision. Interestingly, this also happened in North Dakota. https:
768.
by currymj 8y ago
another neat example of how assumptions about noise and functional forms can let you do causal inference is in an exercise in "Elements of Causal Inference": consider a linear model. The true model is Y ~ aX + ϵ, X causes Y. you w
769.
by currymj 8y ago
I am a happy Julia user, but I can imagine if you had a use case where you wanted to compile a binary or shared library, Julia could be a pain. C++ of course works fine for this but I imagine Swift would be less terrifying to use.
770.
by currymj 8y ago
I cannot tell if this is intended as a very dry joke or not.
771.
by currymj 8y ago
If you can come up with a compelling story and a reason (other than bragging rights), and can prove you have academic skills, you will have a shot. The goal is to convince the admissions committee who are just a few people, so you want to t
772.
by currymj 8y ago
Depends on the goal. If the goal is to maximize publications in good venues and have the occasional spectacular result, I think you are probably right. However, if I grant that there is a lot of low-hanging fruit out there in the economy th
773.
by currymj 8y ago
If you've ever used "Bayesian optimization" to choose hyperparameters, that was almost definitely a Gaussian process. The article focused on the case where you have a finite number of test points, which is probably a good ide
774.
by currymj 8y ago
As others have said, you don't actually want the global optimum of a neural network because that would be terrible overfitting. There is some evidence that architectural tricks (like ResNet) that empirically help performance are making
775.
by currymj 8y ago
most articles from after about 2000 have a version on arxiv, if the conferences and venues aren't already open access. in CS this is almost always substantially the same work as gets published in a conference or journal. Google Scholar
776.
by currymj 8y ago
this is unfortunately almost always true, you just have to try every possible combination of everything with lots of hyperparameters. nothing makes any sense, it's total chaos, and we are wandering blind in the wastelands.
777.
by currymj 8y ago
one would hope so! the depressing thing about machine learning to me is the many convincing explanations of phenomena, that then turn out not to explain things (all the explanations of why dropout is effective, or why there are adversarial
778.
by currymj 8y ago
You can use the TAGME API to do that, it’s state of the art as far as I know. If you want to implement it yourself I think it’s pretty painful, though. No reason you can’t fine tune these on your task, that’s true of any word embeddings not
779.
by currymj 8y ago
i think you’re supposed to click on the links which open up more info written in standard font. in that context, the italics is sort of like how you might get a summary in italics at the start of a book chapter.
780.
by currymj 8y ago
I really think we're talking past each other, and that you must have some more restrictive definition of "algorithm" than me, because I'm reading that page and it walks you through an example in excruciating detail of ho

More ›