MovieLens is a research site run by GroupLens Research in the computer science department at the University of Minnesota. MovieLens uses collaborative filtering technology to make recommendations of movies that you might enjoy, and helps you avoid the ones that you won't. The predictions you get are personalized to your tastes, which are learned by asking you to rate movies that you have seen before.
The original MovieLens algorithm ("user-user") worked conceptually by generating a correlation coefficient between you and every other user in the database. Then, movies were assigned a score based on their ratings by other users who have a high correlation coefficient. With this algorithm, you can be recommended movies that are either highly rated by users with a similar movie preference, or movies that are rated low by users who dislike all of the movies you like. Now MovieLens uses a tweaked version of a different published algorithm ("item-item").