curuinor

43 ms·

451.
by curuinor 10y ago
In large part, in the CPU section, it _is_ working in numpy. I think of the neural network libraries, Chainer is made by the people who like actually coding the most. I mean, for example, lots of the TensorFlow type checking gets done in Ei
452.
by curuinor 10y ago
Gerd Gigerenzer is the author you want for looking at overfit in human beings. Sworn life enemy of the Kahneman and Tversky collaboration, and vice versa(Kahneman actually put in a few disses in his Nobel prize ceremony aimed pretty much at
453.
by curuinor 10y ago
This is an extremely important question. Empirically, exit visas seem like an extremely good feature for having a fucked-up country (North Korea, USSR, Warsaw Pact, Nazi Germany, Fascist Italy all had them, Saudi Arabia does for foreign wor
454.
by curuinor 10y ago
Rumelhart was the first to note this, I think, although of course he didn't have mobile phones. Implementation looks spiffy, but not spiffier than Han et al's work
455.
by curuinor 10y ago
Rather hilariously, deep frying is already a term of art in ML, of course in a radically different setting. Deep fried convnets ( https://arxiv.org/abs/1412.7149 ).
456.
by curuinor 10y ago
This is less of a pedantic point than might be assumed. The Gaussian depends upon a notion of independence and identical distribution: independence can be dispensed of when one thinks about power laws. If you can soberly state the assumptio
457.
by curuinor 10y ago
It's interesting to think that connectionist theories of language are literally shipping on Google Translate and people still argue about Chomsky. Many of the cognitive scientists were poking about with neural nets too - and they were
458.
by curuinor 10y ago
The article mentions a Jerome Siegel coauthored Journal of Neurosci paper in Dec 26, which can only be this paper, which is actually ca. 2007: https://www.semel.ucla.edu/sites/all/files/07%20hcrt%20nasal...
459.
by curuinor 10y ago
Here's the paper mentioned: https://www.semel.ucla.edu/sites/all/files/07%20hcrt%20nasal... So it should be noted that this is a 2007 paper.
460.
by curuinor 10y ago
Full text of the thing: In what sounds like a dream for millions of tired coffee drinkers, Darpa-funded scientists might have found a drug that will eliminate sleepiness. A nasal spray containing a naturally occurring brain hormone called o
461.
by curuinor 10y ago
Cache: http://webcache.googleusercontent.com/search?q=cache:XUB-Gpd...
462.
by curuinor 10y ago
I've done this with custom tools for a few years. Works pretty well.
463.
by curuinor 10y ago
There was a man I knew who actually did quit Palantir to work at a literal candy store, actually (Rocket Fizz, which mainly sells soda but also candy). Interesting guy.
464.
by curuinor 10y ago
It's actually quite a shallow treatment. That's OK. Presumably, most folks who just want to use the things don't want to know about transience in chaotic attractors or VC-dimension and Radamacher complexity and stuff like tha
465.
Ask HN: Command line only CRM's?
8 points by curuinor 10y ago|4 comments
466.
by curuinor 10y ago
C. Alexander in _Notes on the Synthesis of Form_, rather explicitly puts forth the essential architectural problem in a semiformal form - a constraint satisfaction problem. The closest application of CSP-like problems in places close to sta
467.
by curuinor 10y ago
B. Victor was Caltech EE BS and Berkeley EE MS( http://worrydream.com/#!/cv ). V. Hart did a music BA at Stony Brook, but has lots of mathematics background from many sources( https://www.quora.com/What-is
468.
by curuinor 10y ago
I recall CS221 (Intro to AI) at Stanford was a little less than half asian, half white and the rest other.
469.
TIL there's an asymptotically faster neural net layer (for backpropagation)(github.com)
1 points by curuinor 10y ago|0 comments
470.
by curuinor 10y ago
Social presence is your research keyword. There's a long history of research done which can be pretty much summed up with Cliff Nass's contention that people confuse humans with computers a lot, and transactional status ends up be
471.
by curuinor 10y ago
I found that a cronjob did 80% of the job, actually: the pain point is not in remembering the things about them, it's remembering their existence.
472.
by curuinor 10y ago
I mean, TFA says basically that, with a recognition that you still have to have bits of security. I liked the idea about shipping off people to consult at foreign airports.
473.
by curuinor 10y ago
I was referring to measures of _complexity_, variously and ill-defined, which may or may not have anything to do with KC. FI and VC-dim are, indeed, not examples of things closely related to KC, but they are useful measures of complexity no
474.
by curuinor 10y ago
Computable things include CSSR ( http://bactra.org/CSSR/ ), the considerably more bullshit Lempel-Ziv (yeah, like LZ compression) complexity, etc etc etc etc. Lloyd( http://web.mit.edu/esd.83/www&#x2F
475.
by curuinor 10y ago
Badii and Politi 1997 is fun and interesting reading. Try Lempel-Ziv complexity for some hilarious idiocy and statistical Grassberger complexity for some less idiotic thought. Logical depth. Symbolic dynamics.
476.
by curuinor 10y ago
They claim it's twice as fast as tensorflow, which is not blow-you-out-of-the-water (compare to like 50x speedup from GPU on most places), but it's a solid speedup. It's easily parallelizable on GPU's, or so the claim go
477.
by curuinor 10y ago
If by "college algebra" you mean, abstract or modern algebra (groups, rings, fields, modules, vector space, lattices, etc), I would throw away the modern algebra book (for now: there are some applications of topology and topology-
478.
by curuinor 10y ago
You're not asking about anything about computer vision, if you're asking just about Caffe and Tensorflow. You're not even asking about the whole of machine learning: you're just asking about neural networks. If you haven
479.
by curuinor 10y ago
This seems to be long enough for a statistical attack to narrow things down to ~100 people.
480.
by curuinor 10y ago
Some remarks on Mefi itself: http://metatalk.metafilter.com/24089/You-were-accidentally-p...

More ›