At Nanocrowd we are developing new technology for searching popular multimedia content. Our approach will make it much easier for users to find content that is of interest to them. Our first demonstration of this technology is theatrical-release movies, but the methods we are developing will apply to books, online videos, games, music, or even to art and architecture. Soon we will put up a demo on our website.
Listening to the crowd
We learn about movies by analyzing the extensive user commentary and user reviews of movies that are available online thanks to Web 2.0. We apply semantic, linguistic, and statistical analysis and discover the essence of what people are saying about movies. The result is that we have new insights into what people THINK and FEEL about movies.
Finding your voice in the crowd
We called this technology (and our company) Nanocrowd. The crowd part of our name refers to the notion of "the wisdom of crowds." There is indeed plenty of wisdom in the millions of voices that contribute to the Internet. Yet, none of us runs with crowds. We run with friends, with tribes. Even among trusted friends, we will have disagreements about what we like and certainly about which movies we like.
The truth is: matters of taste and preference are not a matter for crowds.
That is the nano part of our name. We listen to the crowd to figure out the unique character of each movie. Once we understand the wildly diverse and interesting opinions of all those online voices, we compare movies to see how they relate to each other, and we identify collections of movies with common and interesting attributes.
Nanogenres
The first thing user commentary taught us was that movies fall into categories that are very different from the ones movie reviewers use. Movies aren't just "Comedies" or "Romantic comedies." Instead they naturally fall into categories we call nanogenres:
"Beautiful love dreams"
"Sweet predictable romance"
"Prisons and redemption"
"Mental patients and their shrinks"
"The French resistance"
"Ships at sea"
These nanogenres are not invented by editors. They are defined by the language used by viewers to describe movies and how those movies made them feel. When we suggest movies, we use nanogenres as our primary roadmap.
Unbiased ratings
Genres are important, but so are movie ratings. Unfortunately ratings are highly subjective and often intentionally biased. We recognize that when people rate movies online, they aren't always accurate or honest. Instead of trusting user ratings, we actually figure out if people like a movie by reading what they say about it. Although we end up with a rating number, we calculated it by interpreting users' words and their meaning. This way we eliminate the usual bias that you find in peer reviews.
Because both our categorization and rating engines are based on Web 2.0 commentary, they are self-defining and self-repairing. Our technology "listens" to reviewers, so as soon as people write about a new movie, we can discover what other movies it is like and figure out if people liked it or hated it.
Hook up with your nanocrowd
Using our technology, you will be able to find multimedia content you will enjoy, because we hook you up with the opinions of the "nanocrowd" that fits you and your mood of the moment.
posted by Roderic March
Posts
