What Does a Neural Network Actually Do?

Originally posted on Some Thoughts on a Mysterious Universe:

There has been a lot of renewed interest lately in neural networks (NNs) due to their popularity as a model for deep learning architectures (there are non-NN based deep learning approaches based on sum-products networks and support vector machines with deep kernels, among others). Perhaps due to their loose analogy with biological brains, the behavior of neural networks has acquired an almost mystical status. This is compounded by the fact that theoretical analysis of multilayer perceptrons (one of the most common architectures) remains very limited, although the situation is gradually improving. To gain an intuitive understanding of what a learning algorithm does, I usually like to think about its representational power, as this provides insight into what can, if not necessarily what does, happen inside the algorithm to solve a given problem. I will do this here for the case of multilayer perceptrons. By the end…

View original 1,775 more words

Kleinhaus: Cabin progress

Just a heads up that I’m still not dead and the cabin designs are progressing well, albeit a tiny bit slowly, since the last time I posted an actual design update. I’ve since named this whole cabin project “Kleinhaus” (German for “Small house” — I thought it was pretty clever). I moved away from fancy pure-Photoshop footwork to basic drawing and then scanning followed by a Photoshop cleanup. I found this to be a tad easier and I’ve had lots of helpful suggestions from folks since I started this journey. Many thanks to all of you.

2 Sheets done (hopefully)

The following are sheets 3 and 4 of the overall blueprint which detail the foundation. I’m still working on the elevation views and floor plan and the overall shape of the cabin has changed a bit to accommodate some much needed design overhauls. Gone are the weird 21-24-21 joist spacing of the first layout. This time, it’s all 24″ on center joists for simplicity and my own sanity. I did away with starting from estimates and then moving on to measuring and instead started with concrete measurements first. I found that this makes more sense since it’s harder to skew corners or screw up the lengths on paper. That doesn’t mean these are free of errors and you should go ahead and start with these measurements without double-checking!

These are only 1200 x 1623 full size, but the actual image I’m working with is 5620 x 7600 pixels (18.73 x 25.33 inches at 300 pixels/inch) which is a tad too large to post here. I’m going to make the full size blueprints available as a PDF or something as soon as I finish and correct any errors.

As always, these are provided as-is. Always check the measurements first, follow local building codes, there be dragons etc… etc…

Kleinhaus 16x16 Cabin: Sheet 3

Kleinhaus 16×16 Cabin: Sheet 3

Kleinhaus 16x16 Cabin: Sheet 4

Kleinhaus 16×16 Cabin: Sheet 4

My heart is ok, but my eyes are bleeding

Originally posted on Leaf Security Research:

TL;DR: heartbleed is bad, but not world ending. OpenSSL is not any more vulnerable because of its freelists and would still be vulnerable without them.

We felt that there weren’t enough heartbleed write-ups yet, so we wrote another one. Unlike many of the other posts, we are not going to talk about the TLS protocol or why we think the heartbeat extension is pointless. Instead, we are going to focus on the bug itself and more specifically, why sensitive data gets leaked.

First we would like to state that, as far as complexity goes, the heartbleed vulnerability is nothing special, but that doesn’t mean it was easy to find. All bugs are easy to spot after someone else points them out to you. Hindsight is 20/20 after all. Riku, Antti and Matti at Codenomicon and Neel Mehta at Google all independently discovered this bug. Neel was also kind enough to…

View original 2,226 more words

Hacker School banning “feigned surprise” is absolutely brilliant

eksith:

“Feigned surprise” should be banned in any organization that purports to bestow knowledge and build confidence as it does the opposite in both.

Originally posted on Coffee Spoons of Code:

[Since you might wonder while reading this piece what my relationship to Hacker School is: I have no relationship with Hacker School. It has been described to me, and I have devoured the blog. If I made a mistake, let me know.]

The biggest insight I’ve had as a programmer is just how often other programmers are portraying false confidence. My natural approach to problem-solving is Socratic, feeling out different ideas and taking small, well-supported steps. Compare and contrast that with making gigantic pronouncements full of bravado. Writing software is inherently an exercise in managing complexity, which is best done with caution.

The best developers I’ve worked with were willing to admit when they didn’t know something. Of course they could learn quickly. If you meet an arrogant developer who pretends to know everything, be careful. To them, their ego is more important than your software. An insecure person who…

View original 244 more words