megalodon

And here I thought Mozilla was driven by the community. I guess there is a price tag on everything these days.

What?

If any search engine should be the default, at least pick the best or the most privacy friendly. Yahoo is neither. On top of that it is just a relay to Microsoft Bing.

How hard is it to just ask the user which search engine they would like to use on install?

Ever visited one of those free live streaming websites? 20 layers of ads with tiny close buttons, that sometimes themselves trigger redirects. It's the perfect example of how ads make you want to punch your screen.

"We also investigated hierarchical features where the image is overlaid with a grid of cell size c × c and pixels withins each cell are added up. This is same as downsampling the image and using the raw pixels in the downsampled image as features." (p. 3)

http://www.eecs.berkeley.edu/Pubs/TechRpts/2009/EECS-2009-15...

I generated training sets for an OCR project in JavaScript [1] a while ago using a modified version of a captcha generator [2] (practically the same technique mentioned in this article).

[1] https://github.com/mateogianolio/mlp-character-recognition

[2] https://github.com/mateogianolio/mlp-character-recognition/b...

Funny, just read an article today proposing the same feature detection algorithm (the one you called 'grid merge'). Have you tried applying these techniques on scanned/photographed documents?

Creating a layout is not a simple problem, but implementing it is. The reason 'most applications suck' is more likely because of the designer (or lack thereof).

Except you haven't learned much after watching those two and a half hours. I would argue reading this [1] or this [2], while simultaneously mocking up a flexbox grid will give you at least the same amount of knowledge in much less time.

[1] https://developer.mozilla.org/en-US/docs/Web/Guide/CSS/Flexi...

[2] https://css-tricks.com/snippets/css/a-guide-to-flexbox/

Especially when there are pages like this: https://developer.mozilla.org/en-US/docs/Web/Guide/CSS/Flexi...

Already done: http://ov3y.github.io/2048-AI/

If you haven't already, check out Synaptic [1] by Juan Cazala. I use it in many of my ML projects and it contains a lot of useful examples and documentation. I recommend you take a look at the source as well, which is easy enough to understand.

[1] https://github.com/cazala/synaptic

"The GUNMAN project had a major impact on the intelligence community as a whole. It brought about a greater understanding of the thinking and operations in a totalitarian society. The community became more aware of the hostile electronic threat against the U.S. As <redacted> explained, "If any other agency such as CIA or the State Department had discovered the bug, this change would not have occurred because they would not have publicized the incident." NSA, however, briefed all levels of government to warn them of the danger. NSA was not out to assess blame; it took the problem-solving approach." (p. 17)

Are you that scared that it might work? It's a social experiment carried out by the government, not judgement day. Let's see what result it produces before rejecting it.

Surreptitiously racist and full of prejudice. Thank you for leaving our country.

Freedom by Jonathan Franzen

Thanks for the suggestion to measure the image processing time, that could be interesting. To be honest I haven't yet tried Tesseract so I can't make any comparisons.

However, I do believe that converting the network input to character codes (output layer size 4 bits for digits, 8 bits for letters) instead of using a softmax layer (would need 10 bits for digits, 26 bits for letters) is a novel approach that really improves performance.

Yes, the success metrics are measured on the training set.

I just ran a measurement with a separately generated testing set consisting of 52k characters which yielded a success rate of 98.53% on letters a-z. This turns out to be almost exactly the same as the rate presented on the repo page (98.52%). Will upload the result for your perusal as soon as possible.

You will have to browse the synaptic neural network library [1] for an answer to your second question.

[1]: https://github.com/cazala/synaptic

EDIT: A separate testing set is now generated in addition to the training set. I updated the success rates of the examples in the readme accordingly.

I'm planning on adding a JSON config file to be able to tweak the network and add a lot more variation to the generated training set. Meanwhile, if you want to train it with several different fonts you can just add them to the 'fonts' array in captcha.js.

Great analysis, I hope you don't mind my linking to this comment as an explanation of the fractal patterns in the readme?

Thanks.

To answer your first question, I (perhaps naively) assumed that the synaptic library used correct naming for its network prototypes [1].

My implementation contains a few modifications to the one defined as 'Architect.Hopfield' [2], which is why I decided to put it in a separate file. It also helps a visitor to know how the network is defined without needing to browse the source of the synaptic library.

[1]: https://github.com/cazala/synaptic/blob/master/README.md#hop...

[2]: https://github.com/cazala/synaptic/blob/master/src/architect...

Nope, but feel free to do whatever you want with it.

You're right, it's just a practical application.

For those interested, I came across this chapter [1], which is great for learning about hopfield networks.

[1]: http://www.cs.toronto.edu/~mackay/itprnn/ps/506.522.pdf

It sure is an interesting output, which is why I put it under 'observations'. For every color (24-bit binary) the network receives it has to output either black or white, which it does, so I wouldn't call it a failure.

Sorry for the confusion, just changed the labelling to "input" and "output".

Link disappeared for some reason: https://github.com/mateogianolio/hopfield-color-filter

Why fight it? Visiting a technological/intellectual inspiration source definitely goes in my list of positive habits.

Just a heads up; the page header (and footer?) does not render correctly on mobile. The top graph is centered and its data is impossible to read because of label overlapping. Interesting analysis nevertheless.

This. The ongoing arms race between advertisers and adblockers damage content delivery, which is why I'm hoping for a solution which can replace ads in the near future. Maybe Google Contributor is onto something.

I would recommend trying the Reader view in Safari (I'm sure there's similar functionality in other browsers) to read articles without clutter. Works like a charm.

I think it's more interesting to do new things with a language I'm comfortable with, than to do things I'm comfortable with using a new language.

Assuming the languages are comparable, of course.

The high level concepts you refer to are implemented in plain javascript. The general purpose of frameworks is to increase productivity for those who already master the language. You will have a harder time learning if you begin by using abstractions you don't understand.