= The AURA Project Wiki = The amount of content on the web, content such as video, blogs, images, podcasts, news, and music is growing rapidly. Much of this content is in the 'long-tail' and as such is difficult for potential consumers to find. Recommendation technology is key to connecting people with this content. Companies such as Amazon, Netflix and Google view recommendation technology as a core asset that leads to increased sales and more engaged, satisfied customers. Amazon, Netflix and Google use a technique called collaborative filtering to make recommendations. This type of recommendation relies on the 'wisdom of the crowds' to recommend items based on who has shown a preference for an item. These recommendations are typically of the form "Customers who bought X, also bought Y". Companies like Amazon, Netflix and Google can generate very good recommendations because they have a large user base, and can effectively mine the taste data of their users. These companies can, however, have trouble promoting 'long-tail' content due to insufficient taste data. There are many companies that could benefit from incorporating good recommendations into their offerings. However, it is difficult for most companies to make effective recommendations. They lack the critical, deep user taste data needed to make good recommendations; they lack the technical expertise to build a good recommendation system; and they lack the computing infrastructure that is required to handle the large volume of taste data that ultimately needs to be collected in order to generate good recommendations. The AURA Project is an open recommendation system that makes good recommendation possible for everyone: for companies that don't have a user base the size of Amazon, Netflix or Google, and for products that don't have the popularity to make it out of the long tail. Aura is a hybrid recommender system that will combine the best aspects of collaborative filtering with content-based recommendation technology to generate recommendations even without large amounts of user data. Some distinctive features of Aura: * Aura can generate good recommendations even when large user taste data is lacking. * Aura can explain why it is recommending something. * Aura avoids the rich-get-richer phenomenon often seen in collaborative recommenders. * Aura is an open recommender service that provides general recommendation web services that can be used by anyone. * As an open recommender, Aura can aggregate taste data from multiple sources, allowing for cross-domain recommendations. Recommendation will be as important to the next generation Web as search is to today's Web. We want to make sure that every company, no matter how large or how small, can generate good recoommendations for their customers.