MozillaZine

Proposals for Incorporating Machine Learning in Mozilla Firefox

Friday June 18th, 2004

Blake Ross writes: "I will be doing research this summer at Stanford with Professor Andrew Ng about how we can incorporate machine learning into Firefox. We're looking for ideas that will make Firefox 2.0 blow every other browser out of the water. People who come up with the best 3-5 ideas win Gmail accounts, and if we implement your idea you'll be acknowledged in both our paper and in Firefox credits. Your idea will also be appreciated by the millions of people who use Firefox :-). We'll also entertain Thunderbird proposals."


#86 Re: Re: Re: Nuke Anything + Bayesian Filter

by superyooser

Thursday June 24th, 2004 4:01 AM

You are replying to this message

You could have predefined sets of priors profiled for different kinds of users: "ad hater," "protective parent," "no flashing stuff," etc.

As for implementation, I was thinking that you could click into "Train Nuke Mode" and draw shapes around what you like or don't like. You know how a lot of news pages have the article in the center column and cruft everywhere else? Draw the rectangle around the article and "crop" the rest out. All the page elements you excluded get fed to the learning filter as *bad*.

Better yet, look at the Web Developer extension for Firefox and see the Outline menu. It draws colored boxes around block level elements (all), frames, and table elements. In Train Nuke Mode, you could select individual elements easily this way. Perhaps a gesture or hotkey could enable and disable outlining elements. There should also be DOM tooltips in case there are lines all over the place and you can't tell what's what.

This is something that the regular Nuke Anything should have. When it says "Remove this object" it's often not clear what *this* object is. Am I nuking a whole table? or just a tr? or td? A div? Or the span, a (href), or font within the div? Sometimes I accidently nuke the main content of the page.