jbarrow

11 ms·

61.
by jbarrow 12y ago
I believe that training methods for recursive neural networks will be some of the most interesting future research in the field. We've seen some pretty revolutionary techniques come out for feedforward models in the last 30 years, but
62.
Church Numerals(karczmarczuk.users.greyc.fr)
30 points by jbarrow 12y ago|9 comments
63.
by jbarrow 12y ago
I actually visited a German University (TU-Braunschweig) that works with a German car company to research driverless cars. The project that I saw [1] used Lidar as a verification method, but the ultimate aim was to be able to drive without
64.
A Better Way to Learn AngularJS(thinkster.io)
1 points by jbarrow 12y ago|0 comments
65.
by jbarrow 12y ago
For anybody who hasn't, I strongly suggest reading On Intelligence, even if you don't agree with all his views. Hawkins does a fantastic job presenting his framework in an understandable way, and once you finish reading it, you de
66.
The Next Generation of Neural Networks (2007)(youtube.com)
3 points by jbarrow 12y ago|0 comments
67.
Scientific GPU Computing in Go(archive.fosdem.org)
5 points by jbarrow 12y ago|0 comments
68.
by jbarrow 12y ago
Jeff Hawkins is a huge proponent of strong AI, and I strongly encourage anyone interested in the subject to read his 2004 book, On Intelligence. In it he makes several cogent points about the future of true AI, a lot of which hit on points
69.
by jbarrow 12y ago
This is a really neat application! Since seeing the fourier image analysis on HN a couple days ago, I've been looking into in-browser DSP. On my laptop it felt pretty responsive, but I notice that you have a disclaimer not to sing a no
70.
by jbarrow 12y ago
I'm glad to see they've finally made a decision. Harry Reid had expressed some concern [1] that Musk might have been playing Nevada with no real intention of building the Gigafactory there, so it's all well that it panned out
71.
by jbarrow 12y ago
The simple answer is that we no longer use the standard linear perceptrons that McCulloch and Pitts put forth in 1943. The limitations of perceptrons and their training algorithms are well-known and well-studied, and in the last 60 years we
72.
by jbarrow 12y ago
It was the IBM SyNAPSE Chip [1]. It's actually been known since 1988/89 that neural networks can approximate any continuous function [2], but this chapter explains how in a much more intuitive sense. [1] http://www.rese
73.
by jbarrow 12y ago
The service looks interesting, although I can't comment on it as it isn't available in my area. I did, however, want to say the the trend of fixed menu bars on web pages should be carefully considered before you choose to implemen
74.
Why we should encrypt everyone's email(popularmechanics.com)
2 points by jbarrow 12y ago|0 comments
75.
A visual proof that neural nets can compute any function(neuralnetworksanddeeplearning.com)
217 points by jbarrow 12y ago|80 comments
76.
by jbarrow 12y ago
I imagine that even using the FFT and not the "normal" DFT, signals processing done in browser is slow for a lot of people.
77.
Basic Bread Formulas(kingarthurflour.com)
2 points by jbarrow 12y ago|0 comments
78.
by jbarrow 12y ago
There are courses on youtube [1] which do a good job of covering a lot of the theory (disclaimer: I haven't finished the linked-to course, but what I have finished has been informative). If you're interested beyond that, you can l
79.
by jbarrow 12y ago
A good overview of the paper is given by Ran Bi: http://www.kdnuggets.com/2014/06/deep-learning-deep-flaws.ht... And a previous discussion of the results can be found at: https://news.ycombinator.com&#x
80.
by jbarrow 12y ago
I've never used SymPy, but after looking through the tutorial, it looks like a nicer symbolic mathematical library than Theano. I know that the libraries have different focuses; Theano is about speed and tensors, while SymPy appears to
81.
Yoshua Bengio on Deep Learning(nuit-blanche.blogspot.com)
5 points by jbarrow 12y ago|0 comments
82.
by jbarrow 12y ago
I have similar thoughts concerning the coolest (and I am an American). My conclusion was that its appeal is inherently different than the Ouya or Pebble, and that my conceptions of what should and shouldn't be a wildly successful kicks
83.
by jbarrow 12y ago
My friends, roommates, and I use it all the time to settle minor debts, pay rent, and split bills. It's simple enough that we can make quick payments, and because it's built on top of email it inherently keeps transaction historie
84.
by jbarrow 12y ago
Extending HTML5 video is actually really interesting. For an old project I wrote a subtitle player for HTML5 videos as a plugin. [1] You pass in a standard SRT file and it automatically parses out the times and changes the subtitle text. [1
85.
Now I Know: A Daily Newsletter of Interesting Things(nowiknow.com)
1 points by jbarrow 12y ago|0 comments
86.
Will astronauts be the norm in 50 years?(virgin.com)
1 points by jbarrow 12y ago|0 comments
87.
by jbarrow 12y ago
Although it's not a discussion of the specific limitations of a closed-world system, Jeff Hawkin's 2004 book "On Intelligence" touches on many of the shortcomings of such systems, especially with respect to general probl
88.
by jbarrow 12y ago
This looks like it has the potential to be a great resource! Not to mention it's coming out of one of the big name schools of deep learning (along with UToronto and Stanford). I imagine, though, that anyone not well versed in college m
89.
by jbarrow 12y ago
For those of you on HN who haven't seen it yet, I strongly encourage watching Doug Engelbart's "Mother of all Demos" [1]. You'll be amazed by what technology was in development in 1968. Keyset was just one part of
90.
by jbarrow 12y ago
I used the SEP when reading about Epistemology, and I must say that it is both a comprehensive and comprehensible resource. Although it was (for me, at least) a bit dense at times, it was perhaps the best introduction to the subject that I

More ›