I disagree.
The people who can most benefit from really accurate, automatic categorization and recommendation of music are not the major labels / FM radio types. They had been getting by just fine with the old system of payola, record stores and using image as a surrogate for talent. The people who stand to really benefit from this are musicians who are making the type of music that YOU really want to hear but that, without automatic indexing and recommendation, you'd never discover on your own. Imagine you ask for "something like Radiohead but a bit more mellow" and, instead of being force-fed more Coldplay, you got a recommendation of a local band who were playing a gig just down the road tomorrow night.
It's true that music is always going to be subjective and that even the best recommenders (e.g., friends or DJs) will make bad choices but with automatic analysis methods, there is the chance to analyze both the music and the listener to help people make really "personalized" discoveries that go beyond collaborative filtering or mass media.
There's a similar type of game at:
http://www.likebetter.com/
This game uses your selection of images to infer facts about you, based on facts about others who like similar images. Pretty spooky...
One complaint about this work is that it requires tagging an initial set of images that are needed to train the system. Vasconcelos' work uses the academic standard "Corel" dataset of labeled images but also uses tagged images from Flickr to train the system.
Using human computation games like the ESP game for images and ListenGame www.listengame.orgfor audio, collecting data is not as tough as it once was...
Tagging images (as this system does) is in fact an effective way to do "real" image search. The work in Vasconcelos' lab and similar work on music annotation/retrieval UCSD's Computer Audition lab. has shown that representing an image as a distribution over semantic concepts (basically a "tag cloud") makes more reliable and accurate searches than using image- or audio-based feature comparisons alone. Basically, it's easier to find an image if you tell the computer "I'm looking for a white, fluffy cat" than if you say "I'm looking for a connected, amorphous set of white-ish pixels with a fluffy texture".
Slashdot Mirror
I disagree. The people who can most benefit from really accurate, automatic categorization and recommendation of music are not the major labels / FM radio types. They had been getting by just fine with the old system of payola, record stores and using image as a surrogate for talent. The people who stand to really benefit from this are musicians who are making the type of music that YOU really want to hear but that, without automatic indexing and recommendation, you'd never discover on your own. Imagine you ask for "something like Radiohead but a bit more mellow" and, instead of being force-fed more Coldplay, you got a recommendation of a local band who were playing a gig just down the road tomorrow night. It's true that music is always going to be subjective and that even the best recommenders (e.g., friends or DJs) will make bad choices but with automatic analysis methods, there is the chance to analyze both the music and the listener to help people make really "personalized" discoveries that go beyond collaborative filtering or mass media.
There's a similar type of game at: http://www.likebetter.com/ This game uses your selection of images to infer facts about you, based on facts about others who like similar images. Pretty spooky...
One complaint about this work is that it requires tagging an initial set of images that are needed to train the system. Vasconcelos' work uses the academic standard "Corel" dataset of labeled images but also uses tagged images from Flickr to train the system. Using human computation games like the ESP game for images and ListenGame www.listengame.orgfor audio, collecting data is not as tough as it once was...
Tagging images (as this system does) is in fact an effective way to do "real" image search. The work in Vasconcelos' lab and similar work on music annotation/retrieval UCSD's Computer Audition lab. has shown that representing an image as a distribution over semantic concepts (basically a "tag cloud") makes more reliable and accurate searches than using image- or audio-based feature comparisons alone. Basically, it's easier to find an image if you tell the computer "I'm looking for a white, fluffy cat" than if you say "I'm looking for a connected, amorphous set of white-ish pixels with a fluffy texture".