currymj

71 ms·

811.
by currymj 8y ago
I think the idea of Pop is that it's for people like scientists and engineers who need to use their machine as a tool for computation, but may have no knowledge or interest in system administration. Likely quite comfortable on the comm
812.
by currymj 8y ago
i mean, probably the sexiest application of linear algebra is machine learning and different exotic norms are very important there. sometimes you even have to learn a norm from data (metric learning). at the very least i feel like it’s impo
813.
by currymj 8y ago
I would also add two good reasons to stick with R: RStudio and Hadley Wickham. In theory, there are Python and Julia equivalents to RStudio (JupyterLab, Spyder, PyCharm, Juno, whatever) but RStudio is just so, so, so good. A truly great pie
814.
by currymj 8y ago
I hate to be the stereotypical Julia recommender, but it is made for this use case, more so than Python, which isn’t all that much faster than R if speed matters. (Unless you want to try Cython but that’s a whole bag of worms.)
815.
by currymj 8y ago
The original IPython notebook was consciously imitating Mathematica. I don’t think it quite makes sense to compare these notebooks to Knuth’s literate programming. The whole point of that was that you could present things out of order, whic
816.
by currymj 8y ago
You are right that they are very closely related. For variational autoencoders (one baseline technique for this sort of thing) if you make your neural network have one layer without a nonlinearity, and train it, it ends up minimizing the sa
817.
by currymj 8y ago
aesthetically i sort of agree, but i just want to type in equations from papers without having to do any error prone arithmetic — it is slightly less frustrating.
818.
by currymj 8y ago
surprisingly yes, it really is. Julia is smart about type inference, so it compiles a specialized version of a function as needed for whatever argument types actually get used. the catch is, the first time you run a function there is often
819.
by currymj 8y ago
right, isolated small groups or individuals believe the fake videos (or "believe" them, it's possible some of these groups were already looking for an excuse for violence). but it's not really the public sphere -- there
820.
by currymj 8y ago
See Dai et al, "Learning Combinatorial Optimization Algorithms Over Graphs", experiment section A.1.3 "Traveling Salesman Problem". https://www.cc.gatech.edu/~lsong/papers/arxiv_rl_combopt.pdf
821.
by currymj 8y ago
its not always true (hash functions are in some sense designed by brilliant people to be as hard to learn as possible) but it’s true often enough to be surprising. if there was some score that would look at lighting etc, you probably could
822.
by currymj 8y ago
i also can’t think of a case where a forgery has caused a serious problem in the public sphere. however, there are tons of gullible and/or wilfully stupid people who respond to photoshopped pictures on social media as if they are real.
823.
by currymj 8y ago
While having a Torch-esque GPU ndarray is great, the ability to easily write your own kernels without having to compile gnarly C++ code is what sets Julia apart from competitors IMO. Not sure if there's any other dynamic language offer
824.
by currymj 8y ago
the contract might not be literally enforceable by the courts, but the real threat of costly litigation can both intimidate employees and companies that might hire them.
825.
by currymj 8y ago
80,000 hours appeals to a particular type of person, and I think this is good but vague advice for that sort of person. Reputation and social standing matters, you have to network, be humble and learn from people older and wiser than you, d
826.
by currymj 8y ago
not turned on by default but they are sort of a default choice for vision tasks. it may also be that they were useful on an earlier iteration of the model, and only became harmful later. nobody’s quite sure why anything works, though. hence
827.
by currymj 8y ago
I've seen "hacker" in the non-security sense translated as "魔法师 / mofashi", which literally means "magician" or "wizard".
828.
by currymj 8y ago
I would add, as a compromise with the superintelligence true believers, "Concrete Problems in AI Safety" is quite good, and worth reading and thinking about. (It's especially relevant to RL systems that interact with the worl
829.
by currymj 8y ago
I do not think you should be worrying about superintelligence. If you are investing in AI firms that are going to make products that will be profitable, providing a return on investment, they should focus on making the products, not on maki
830.
by currymj 8y ago
I recommend the book “Paradoxes in Probability Theory”, by William Eckhardt. It refutes a number of these thought experiments that play weird games with probabilities, including the Doomsday Argument and the Simulation Hypothesis. I find it
831.
by currymj 8y ago
Can someone comment on whether these are respectable journals or not, within their fields? Every few years someone will do a Sokal-hoax-esque thing in scientific fields, but usually they're just getting a paper into very-low-quality pr
832.
by currymj 8y ago
I get the complaint about credentialism in certain degree programs and the worry that it will spread into what I assume is your own field. But the comparison of human beings to cockroaches is very troubling to me; it sounds much harsher tha
833.
by currymj 8y ago
I mean, technically, they are a manufacturer of Linux boxes (the DGX Station has Ubuntu preinstalled), so intuitively one might expect they would offer good Linux support.
834.
by currymj 8y ago
The tl;dr is it sort of lets you write a program that describes sampling from your data-generating model, perhaps setting priors over certain unknown parameters. by allowing for programming language constructs, you can define very rich and
835.
by currymj 8y ago
Key distinction about "ban the box", it doesn't make it illegal to ask employees about a criminal record, or refuse to hire them if they have one. You just can't ask on the application form. You can ask later on in the p
836.
by currymj 8y ago
My understanding is that arbitrators do actually rule against the company from time to time, it's not a pure rubber stamp. It seems like one thing to do would be to make sure to rule against the company for small cases with limited dam
837.
by currymj 8y ago
So am I right in concluding that the design was simple enough to be mostly forking-paths-free, and there weren't any oddly specific choices made in the analysis (to dig a "significant" result out of the data)? The fact that i
838.
by currymj 8y ago
Given the journal and the time it was published, are we sure this is research that would replicate? I looked at it and I don't see anything egregiously bad-looking, but I don't know much of anything about experimental design. I ha
839.
by currymj 8y ago
Probably not; the GAN technique wasn't even published until 2014, and this is based on other work that's been done even more recently. The field moves fast and is more open than you might expect (because it's still so academi
840.
by currymj 8y ago
I thought he was more suggesting that the proponents could bolster their case by using their weird mathematical machinery to prove some other, less daunting, results or otherwise show that it has more general uses. (Not asserting that this

More ›