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."
#56 Sidebar Viewer
Monday June 21st, 2004 6:30 AM
You are replying to this message
Most of the obvious and best suggestions have been made: 1) Learned prefetching, either based on relative/absolute linking or learned through link histories and page location. 2) Smart bookmarking, including a cycling top 10/25/50/n most visited links. 3) 'Related To' features that find pages similar to current page. 4) Content filtering based on mouse-selection profiles, either whitelisting or blacklisting. I just wanted to point these ideas out as confirmation that they are relevant.
Unrelated to Machine Learning itself, whether it be prefetching or content filtering, Firefox would do well to have a sidebar that would display the page title, url, and percentage of prefetch or modified construction complete. Additionally, and more importantly, it would have a toggled parameter that would allow/disallow the display of a thumbnail of the page as it would look displayed in the main window. When the mouse cursor moved over these thumbnails, either an enlarged view would raise up or the page would temporarily be displayed in the main window - that is until the cursor was removed.
In the case of prefetching, the user could tell if the prefetch was useful before s/he visited the link or based on a confirming click-over to the main window. With content filtering, the sidebar could display multiple guesses at what the user would like to see or not see. These thumbnails could then be whitelisted, blacklisted smartly to determine future prefetches or content-filtering for individual domains. There should also be a UI, whether a sidebar or dialog, that allowed the user to delete site profiles so that excess learned data can easily be discarded.