Machine Learning (Theory)

4/18/2007

$50K Spock Challenge

Tags: Competitions,Machine Learning jl@ 5:24 pm

Apparently, the company Spock is setting up a $50k entity resolution challenge. $50k is much less than the Netflix challenge, but it’s effectively the same as Netflix until someone reaches 10%. It’s also nice that the Spock challenge has a short duration. The (visible) test set is of size 25k and the training set has size 75k.

5 Comments to “$50K Spock Challenge”
  1. eric says:

    I got really excited there was another contest in the area, however I cannot figure out what value this company adds. There are not only better problems for smart people to be solving instead of essentially donating to a product that seems like a feature limited and data limited wikipedia. How did this company get $7 million?

  2. Brian says:

    Check out the video at: http://news.com.com/1606-2_3-6177391.html?tag=ne.video.6176053&tag=cnet.sc

    It’s actually a pretty compelling product.

  3. Wayne Kao says:

    Although it’s what’s highlighted in many demos, Spock goes far beyond searching the limited set of Wikipedia people. We have millions of people indexed from various online sources including social networks, so even normal people like myself are in there. Wikipedia only hopes to include all celebrities; Spock’s goal is to search everyone.

    The leads to the problem of figuring out which of the many Internet data sources correspond to the same person, a non-trivial problem we hope researchers can help us with by taking on the Spock Challenge. That way you can go to one place and find all the stuff online about that person. Like John wrote, we’ve structured the challenge so researchers have something to shoot for in the near term, with essentially an equivalent reward.

    You can see a short video demo of Spock at the end of this CNET video: http://news.com.com/1606-2_3-6177391.html

    –Wayne, a developer at Spock

  4. hal says:

    FWIW, you can do remarkably well on this sort of task using largely unsupervised approaches (eg., this paper); though maybe if there are 5000 “john smiths” that you’re trying to separate, things get harder.

  5. [...] The spock challenge for named entity recognition was won by Berno Stein, Sven Eissen, Tino Rub, Hagen Tonnies, Christof Braeutigam, and Martin Potthast. [...]

Leave a Reply


Powered by WordPress