florianletsch

(EDIT due to wrongly stating that models run on the Raspberry Pi directly)

The Google Vision Kit will run models on a custom neural processing chip connected to the Raspberry Pi Zero. With the DIY setup from the blog bost, the neural network runs on a "large pc" (potentially with GPU). Depending on the hardware you have at your disposal, you can run more complex (and therefore more powerful) neural networks. At the same time, you'll need wifi set-up and streaming to work. Completely embedded devices are easier to just put in the wild.

In theory, you should be able to use the models from the Vision Kit if you follow their instructions and just put the on a Raspberry Pi directly, and get an additional Movidius compute stick: https://developer.movidius.com/

Where can I follow your project? I currently use QOwnNotes on Linux which does exactly that, but would be keen to try some alternatives, in particular if you're working on a mobile version.

> (...) the current hype about AI replacing jobs has much more to do with media clickbait than any real issue we should be worrying about right now.

Hype? For sure. But the current advances in autonomous vehicles will have a very real impact on the working life of the common truck driver (3.5 million in US alone), I think we all agree on that.

Care to share some details on your technology? On whose handwriting was this trained, did you use any public datasets for this? And of course, how well will this perform on writing styles it hasn't seen before?

Maybe just browse the right subreddits, start posting your own content and then get in touch with people? https://www.reddit.com/r/gameDevClassifieds/ https://www.reddit.com/r/gamedev/ https://www.reddit.com/r/PixelArt/ etc

You probably know this already, but if you are just looking for an illustrated demonstration of what RNNs are capable of in the text domain, probably the best brief article is the post by Andrej Karpathy about [The Unreasonable Effectiveness of Recurrent Neural Networks](http://karpathy.github.io/2015/05/21/rnn-effectiveness/).

Sidenote: A common practice is to take a pre-trained model (e.g. on the Imagenet dataset) and only learn the last few layers for your usecase. This way you can get a well trained feature extractor if your task data is similar, and then only train the classification, which is a lot faster than full end-to-end training.

You can call it an artificial neural network (ANN) as soon as you have two or more artifical neurons connected to each other. As simple as that.

A small amount of neurons might already solve some problem you're having. The XOR problem can be learned by 4 neurons connected to each other.

When you want raw images or similar as an input and have it be classified into 100 classes (e.g. look up CIFAR-10 or CIFAR-100), you will need an architecture with many more neurons.

After all, ANN are simply a tool. Depending on the task, that tool might need to more elaborate. And when you have all those different possible architectures, you want a common way of naming them. Labels such as Deep Learning are simply nomenclature of talking about certain groups of artificial neural networks.

Have been running the trial for a few hours and am really satisfied so far. I see the same issues with Spectactle, though. Hope this can be addressed in a future version, as I'd love to have both apps running permanently.

Still building up a user base so we can have better results, but here is the top list of our platform: https://podcastprofile.com/top (the project is in its early stages, basically the result of a hackathon we had a few weeks back)

It should be "a computer beating the human" in my opinion. It's just like parents bathing in the success of their offspring. Yes, they can be proud, but no, they didn't achieve the "win" themselves.

MacPass is excellent. It finally helped me completely move to a password manager based life.

The submission title is misleading.

> NOTE: This is a DRAFT of the Go 1.6 release notes, prepared for the Go 1.6 beta. Go 1.6 has NOT yet been released. By our regular schedule, it is expected some time in February 2016.

Not based on human subject's opinion but on measured performance:

> The images had each received a “memorability score” based on the ability of human subjects to remember them in online experiments.

Are Facebook access tokens vulnerable to such an attack?

If I remember correctly, at least Google tokens wouldn't be: The application receives a token from Google. With that token, a new session token is created. This session token expires and can only be renewed with the application token and the correct redirect URL.

If Facebook uses a similar scheme, tokens would be useless without the running application renewing session tokens.

Veritaseum did a nice analysis and summary of this exact situation. https://www.youtube.com/watch?v=oVfHeWTKjag

Hm, why do most of his files end with 10 to 15 blank lines of whitespace? Is that a thing or was he just sloppy?

Not quite, but close http://i.imgur.com/dt3hEYc.png

I don't know what to make of this. I like the general idea but I really don't like the use of Markdown here.

Markdown makes writing markup faster and reading markup easier. It's a limited set of syntax elements optimised for writing and reading text. No more, no less. It is not, however, a new set of language elements to be used for whatever purpose.

For example, how is this:

  __Basic Button|Icon：User,Text：Add Friend__ 

Any better than this?

  <button icon="user" text="Add Friend">

Even YAML seems to be much better suited for structured data like this.

  button
      icon: text
      text: Add Friend

If your aim really is to "lessen the cognitive load", why not work with tools that better work towards that goal?

The article list reads like what Hacker News would have been in 1977.

Why so many computer languages?

The hobbyists operating system

The 8080 you may have missed

...

If you're looking for something modular and customizable, be sure to checkout http://getuikit.com/ as well.

The "cyberwar" which you're undoubtedly referring to is one thing. Mass surveillance of the world's population is something else.

Yes, the future is digital. And yes, there will be dangers from several global parties. But why does the NSA chose to effectively weaken public communication by discovering and using security issues in known protocols instead of making sure those get fixed (if the NSA can find it, who's to say no one else can?).

With their proven effort to make everyone's communication interceptable (and effectively weakening crypto systems in all possible manners) they aren't protecting anyone, but instead putting US companies, US citizens and all global users of the web at an even higher risk.

I haven't even started on the effects on civil rights and freedom of speech of a global surveillance apparatus that is acting in secret and not under democratic control.

If it would be about an arms race, why not enforce secure crypto standards and help the industry in that regard? Clearly it's not about having more "cyber power" than China, North Korea and what have you. Instead, it's about having power on individuals, no matter where they're from.

NZ Prime Minister John Key in August, 2013: "I'll resign if GCSB conducts mass surveillance". Well, now'd be a good time.

Source: http://www.nzherald.co.nz/nz/news/article.cfm?c_id=1&objecti...

I love Project Euler. In this thread I've found references for similar projects but focussed more on actual coding (http:(http://oj.leetcode.com) or on BioInformatics (http://rosalind.info/problems/locations/).

Any chance you guys know of a similar project for Computer Vision or general AI topics?

> 101. Don't worry. You'll survive. It's Netscape we should really worry about.

Why did no one listen?