Wouldn’t it be great if instead of spending hours mining feeds and Twitter each day you could just go to a web page and have a personalized stream of news? A stream that learns about you over time and continues to get smarter about giving you just the right news for you.
Yesterday at the Neural Information Processing Systems Conference in Vancouver, Ralf Herbrich and Jurgen van Gael of FUSE Labs showed the latest version of Project Emporia – it goes a long way towards providing this kind of service. This research project has come a long way since it was first shown at the Thinking Digital conference in the UK earlier this year – an all new HTML5 UI for starters.
Project Emporia is a recommendation engine for news. Based on the Matchbox technology from Microsoft Research, it uses a Bayesian probabilistic model to learn the preferences of users for recent news stories.
In plain English, that means Emporia recommends news for you based on news you have read earlier – it predicts what you may want to read. When using Emporia you can vote a link up or down and that influences what you continue to see. From a personal point of view, I wish someone would do this for television and stop showing me ads for stuff I really don’t need :) In fact there is a whole bunch of things I’d love to see this applied to but that’s the topic for another post.
Emporia mines RSS feeds and all links shared on Twitter – discovering around 1,000,000 articles every day. Not only that, it does automatic classification of articles into categories using another Microsoft Research technology for classification. Of course that’s not enough for Ralf and team so they’ve developed a system for “active learning” to automatically discover links that cannot reliably be classified. They sit down and do this manually each night by hand. Just kidding - these types of links are automatically sent to Amazon Mechanical Turk for labeling, they’re then “spam filtered” and returned the classification model to be appropriately categorized.
Fortunately all of that big brain stuff happens behind the scenes leaving the user with an elegant interface that presents personalized news.
I got to play with the app this week and I can report that it works very well indeed. You can see the demo from NIPS at http://emporianips.cloudapp.net/#/topStories - stay tuned for release and some more exciting news I have coming,
what is the difference between this and montage?
@scout the difference between Montage and Emporia is that Montage is a tool to create web pages while Emporia is a service or site that delivers personalized information and learns about what you like to make the content more personalized over time
Veveo had very product to recommend news videos based on your past history