Given so many webpages have dynamic elements, piece of HTML/DOM that only pop up in edge cases (alerts like cookie notices, loading animations, confirmation dialogs) I would be very wary that a tool would be able to identify all the things that are really in use. To use effectively, I think you would need a test script that stepped through all the edge cases, and then you're depending on your test coverage being 100% (of at least triggering the possible DOM structures and style class configurations).
I have the following URL as the first bookmark on all browser bookmark bars:
data:text/html, <html contenteditable>
It isn't nearly as featureful as a full Markdown editor, and it won't save the contents of the tab to localstorage or anywhere else, but it is about as simple a solution as you can get.
I appreciate the "dots don't matter" feature, but have always felt it required an API for external developers to canonize gmail addresses for this precise reason.
If that was the goal, wouldn't efforts be better placed in the Jr Colleges, online courses, and similar institutions. Grade-school level curriculum is a pretty far-future return on investment is the only goal is cheap workers.
Our goal is to reduce barriers surrounding syntax and discoverability, so that they reach computational skills (sequences, conditions, functions, lists and other data structures) faster and with higher confidence and motivation. With the motivation (e.g., "I already built this") and a solid foundation, text programming can feel more achievable to the student.
Blocky team member, here. Microsoft and MIT/Scratch have been amazing partners and contributors. We're really excited about Scratch 3.0 and the efforts on Scratch Blocks.
I don't do neural nets, but if I had to crudely estimate...
10,000 songs (the outputs) * 16 layers * 16 parameters per node * 4 for the bytes per float = 10MB + a DB of song/artist names.
I'm probably underestimating the parameters per node, and overestimating the size of the layers closer to the input. Further, it's more likely structured as an LSTM than a convolutional network, since sound is a streaming source.
Ummm... isn't that a pretty major effect? Doesn't that have some impressive side effects?
I watch my so many of my non-CS friends cycle through jobs (and debt) with limited ability to add long-term stability. If information sciences are having an impact on every industry, and encouraging students to explore those options gives them better access to jobs, and thus improve their life, that seems like a noble goal.
I'm part of the Blockly team (https://developers.google.com/blockly/), an open source project for visual drag-and-drop programming, usually targeting kids. Despite being a "visual programming editor" first, we are exploring blind accessible (i.e., screen reader ready) variants of our library.
Right now, it is effectively a different renderer for the same abstract syntax tree. We'd love to see people evaluate the direction we are currently going, and possibly apply the same accessible navigation to our existing render.
In terms of dev tools, Blockly blocks are usually constructed using Blockly (https://blockly-demo.appspot.com/static/demos/blockfactory/i...). That said, no one has considered what it would take to make our dev tools blind accessible. The fundamentals are there.
Granted, Blockly programming is far from being as powerful as other languages. It is aimed at novice programmers, whether for casual use or to teach the fundamentals of computational thinking. You can write an app in Blockly (http://appinventor.mit.edu/).
My gut thinks this sort of training alternative to back propagation has a lot of uses where SVM have no applicability. The article talks a lot about RNNs (neural nets for sequence prediction), but I would guess it would have uses in online learning as well. Learning twice as fast in those situations seems pretty significant to me.
The computer does not duplicate selective detail. Artists will put detail in areas of focus, and intentionally obscure or abstract other areas. Also, there is often an interplay between medium and message. To that last point, I'm guessing there is still a human "artist" who chose the reference scene and pose and artist style, even if they didn't paint it directly. Towards that, I see this as a creative tool to be included in some future Photoshop, or similar.
Also... Share on Facebook? "App Not Setup: This app is still in development mode, and you don't have access to it. Switch to a registered test user or ask an app admin for permissions."
https://www.cs.ox.ac.uk/ss2014/programme/Bob.pdf