Dn_Ab

47 ms·

271.
by Dn_Ab 14y ago
Actually Nemerle is a language worth talking about. What those guys are doing is brilliant. I remember even years ago - playing with the language - I was able to embed a natural syntax for a constraint programming library, encode a clean sy
272.
by Dn_Ab 14y ago
It depends on what you want to do. It would make sense that Scala, a language that can naturally encode a type class like structure and jruby, a dynamic language wouldn't mind erasure. For jruby it is self evident why and for scala's case,
273.
by Dn_Ab 14y ago
Ivo's answered so I'll just say the main trick to turning a function tail recursive is to have the function lug its state around. So always look to see if you can rewrite your function so it passes the current value explicitly to the next i
274.
by Dn_Ab 14y ago
You got the core, I will correct your edges. The [1..3] is just sugar for 1:2:3:[] Product is not actually an abstraction as it's restricted to work on number types. And no foldl is not an abstraction of (2) because foldl is tail recursive
275.
by Dn_Ab 14y ago
I lament the overlarge emphasis on entertainment careers by the general populace as much as anyone who will lament such things but. Humanity must progress more than technologically, it must also progress culturally. What point is there to c
276.
by Dn_Ab 14y ago
Does exercise induce autophagy in neurons? What are the most interesting developments in research?
277.
by Dn_Ab 14y ago
Vinge did coin the term singularity. The term was coined by science fiction writer Vernor Vinge, who argues that artificial intelligence, human biological enhancement or brain-computer interfaces could be possible causes of the singularity
278.
by Dn_Ab 14y ago
1.) Okay. 2.) Singularity as in breakdown not as in single. You purposely muddled the meaning to make your quip work. 3.) Moore's law fueled as in AI gets interest on their intelligence. Recursive as in AI makes smarter AI makes smarter AI.
279.
by Dn_Ab 14y ago
The main characters of both papers are many layered neural network architectures, autoencoders and stochastic gradient descent. The interesting thing is that all these ideas are from the 80s but the breakthrough was in how to use unsupervis
280.
by Dn_Ab 14y ago
Vinge is a retired math professor, I.J. Good was an accomplished mathematician, Kurzweil makes hard to swallow predictions but is still an accomplished technologist and I happen to very much enjoy Vinge's writing. Regardless, the idea is wo
281.
by Dn_Ab 14y ago
For those keeping score, google's image feature extractor shares the same core principles as microsoft's speech recognizer. EDIT: by keeping score I mean keeping track of which techniques are being used where.
282.
by Dn_Ab 14y ago
Your characterization of Vinge's singularity is incorrect. I have never read anything in which he brings up infinity, I.J. Good does though. Vinge's is actually more like your AI revolution except that it will be evident as a singularity o
283.
by Dn_Ab 14y ago
>> To improve CV we should focus on finding best ways of converting full resolution visual data to something of smaller volume in such way that important features are preserved. You are exactly right! See Dictionary Learning, Rand
284.
by Dn_Ab 14y ago
The article authors of the paper? How can you say they are deeply confused - have you not seen their previous work and presentations? Everything else you say I agree with.
285.
by Dn_Ab 14y ago
The performance is due more to the architecture than to the scale. The scale is to handle all that data. But the feature learning performance has to do with their layered sparse learning technique which is brilliant. Although, their autoenc
286.
by Dn_Ab 14y ago
If you are trying to make game AI then neural networks are a bad idea. But if you are just trying to have some fun learning then do whatever you want. That said, MLPs are hard to use properly, not a good place to start. A good way to start
287.
by Dn_Ab 14y ago
A human baby learns in a very interesting way. By using an overcomplete basis to sparsely code data. This method has only recently started getting attention in ML. A human baby learns from uncleaned raw data using far less energy with bette
288.
by Dn_Ab 14y ago
To the contrary I think people are too eager to find scale free and power law phenomena. There are several papers discussing log-normal bulk and then power law tail. http://physicsworld.com/cws/article/news/2002/feb/26/the-phy... http:/
289.
by Dn_Ab 14y ago
As far as I'm aware the debate of lognormal vs power law distribution on wealth is unsettled. EDIT: google search suggests power law tail - log normal bulk.
290.
by Dn_Ab 14y ago
And so we get a lognormal distribution where small positive differences (arguably ultimately due to luck; genes, environment, continent, family network etc) combine to give enourmous advantages. That said, regardless of your starting point
291.
by Dn_Ab 14y ago
QuickCheck is great and can be found in many languages - dynamic and static alike - nowadays. Another tool that is currently unique to Haskell and is easier to write tests for is Lazy/SmallCheck. SmallCheck shines in the left long tail of f
292.
by Dn_Ab 14y ago
OCaml's nephew, F# does have bounded polymorphism though. Although it inherits that from the OOP side of its family.
293.
by Dn_Ab 14y ago
Using Java and C# as an example of typed languages is like using Youtube comments as a representation of humanity. Once you get a hang of them, writing in any modern statically typed declarative language enters you into a state such that co
294.
by Dn_Ab 14y ago
Heh I had to parse your sentence twice as (+) is a function in Haskell.
295.
by Dn_Ab 14y ago
This is nothing like failing to recognize that Laplace did most of the work in Bayesian Probability. Check http://en.wikipedia.org/wiki/Memristor#Historical_timeline . The subject did not suddenly explode from nascent form to real subject
296.
by Dn_Ab 14y ago
Perhaps I've been warped by machine learning but it seems obvious to me that labeled color differences would be easier to distinguish. Since each time we used the word we would be training our brain to discriminate between color input patte
297.
by Dn_Ab 14y ago
I think Alan Kay is best interpreted as saying "there has not been a Moore's law level of progress for software", not "there can be no Moore's Law for software". I can think of two metrics by which software progress can be measured. Measure
298.
by Dn_Ab 14y ago
There must be a reason why they are confident current algorithms will have a hard time. The article says it works by playing speech backwards so any basic filter will also significantly distort the speech part since I assume the "noise" wil
299.
by Dn_Ab 14y ago
Am I the only one who doesn't use bookmarks? No. I do, but there are a lot of people with different approaches, so all that can be said is the number who bookmark is large and the number who don't is also large. What does everyone else u
300.
by Dn_Ab 14y ago
I remember you said once that rote practice and drilling of arithmetic helped even higher math. I agree. How is drilling mental math any less mechanical than the unconscious familiarization/ingraining of algebraic concepts of this app? In b

More ›