Jesse Vig is a different kind of control freak: He wants the rest of us to have more power to make our own choices -- starting with movies.
So the University of Minnesota computer-science grad student devised Movie Tuner, a cool feature at the movie-recommendation website Movielens, a research project of the University of Minnesota. It allows users to go beyond "if you like that, you'll like this."
"Recommender systems have been around since the '90s," said Vig, 38, of Roseville. "That technology has been perfected; it's all based on artificial intelligence or machine learning. This is a way of combining that sort of machine intelligence with user control."
Once users register at the free site (www.movielens.org), they can look up a film and use Movie Tuner -- which Vig designed for his doctoral program with professors Shilad Sen at Macalester College and John Riedl at the U -- to customize tags that characterize the film. The picks for similar movies change depending on if you want more or less of a given quality.
The tags are a work in progress. For example, one of the tags for "The Social Network" was "serial killer." But the effect can be minimized with a few clicks.
Vig tested the system to great effect. He started with "Mission: Impossible" and added tags for a movie recommendation that was "more realistic, grittier and more intelligent," he said. "You could also say 'less Tom Cruise.'" He discovered what is now one of his favorite films, "The Bourne Identity" -- and no Tom Cruise.
