Next week I'll be representing Brian McFee at AdMIRe 2012 in beautiful Lyon.
And since our paper is called the Million Song Dataset Challenge, well, it is no big secret that a recommendation challenge around the MSD is coming soon.
To be honest, the original hope was to launch something one or two months ago, but you know how it goes...
Anyway, this post is not the official launch, but if you're interested in music recommendation, you might want to read the paper, start thinking about what's right and wrong about the setup, and set aside some time in the upcoming months to try a few submissions of your own.
Congrats to Brian, and don't hesitate to write him if you have questions about the paper. For data-specific issues I might still be the best first contact, and I suggest starting playing with the Taste Profile subset. If you're in Lyon, come say hi! I'll be the one refilling on camembert.
Final note, as usual, the MSD is a community effort, and none of this work would be possible without The Echo Nest, CAL Lab, LabROSA, Kaggle, NEMA, the MSD Challenge advisory committee, ISMIR organizers, and I'm sure I'm forgetting many others.
Cheers everyone!
-TBM
- millionsong's blog
- Login to post comments