kaggle/quora duplicate question contest - CV/SE engagement

Question

There is a kaggle contest for duplicate question detection:
https://www.kaggle.com/c/quora-question-pairs

winner gets a few thousand dollars.

This is a big deal for CV and SE. The CV folks are likely kagglers, and could make some solid steps. It is also a place for some decent visibility.

In what ways is the cross-validated community going to engage this?

• are there teams I should watch?
• will highest CV score be a thread here in meta?
• when the winners blogs on "how I did it" are made, are we going to look at them for use here?

Answer 1

If there is a team from Cross Validated participating, it would be interesting to get a report as an answer to this question.

Obviously, there might be an application for Stack Exchange depending on how the algorithm makes a match. We are very interested in catching duplicate questions for our network too. Our reasons are almost identical to Quora's:

Quora values canonical questions because they provide a better experience to active seekers and writers, and offer more value to both of these groups in the long term.

Additionally, duplicate questions muddy the waters when it comes to search. We'd prefer people to find one question with great answers when querying Google than several questions. It's better for the site, better for people searching for answers and even (in the long run) better for the company.

As an aside, you might be interested in a deduplication project from the University Of Melbourne.