Friday, June 29, 2007

Netflix Prize: The Long Walk

ML@UToronto, a team representing one of the strongest machine learning research groups, regains 3rd place with their new solution.

Netflix Prize Leaderboard - Top 10

5 comments:

Anonymous said...

Your Netflix result is very impressive!

NIPS Reject

Anonymous said...

Check out my new blog on chronicling my adventures on the Netflix Prize http://patternexcavator.wordpress.com/

AP said...

Hello NIPS Reject,

Welcome to my blog and thank you for your positive comment.
Thanks for sharing your team's approach of using Restricted Boltzmann Machines. It was a very interesting reading.


Patternexcavator,

Thanks for the link. I've already found your blog today using http://blogsearch.google.com


Best Regards
-Arek

Tom Slee said...

Congratulations on your progress. I'm not involved in the prize myself, but made some sceptical remarks about it from a different point of view at http://whimsley.typepad.com/whimsley/2007/07/the-limitations.html.

I'd be interested in the opinion of someone who has looked closely at the data whether I'm on the wrong track in my assessment.

I do have doubts about the merits of recommender systems, but not about the skill needed to do well in this kind of competition!

ap said...

Tom,
Thanks for the link to your article. I've read it. Generally, I agree with presented facts and numbers. I can't fully agree with conclusions.
It's not clear to me what are your doubts about. Is your thesis that it's better to browse movies without any order instead of in order presented by a recommender system?