Posted by: Chris Brew | May 10, 2012

Kaggle and automated essay scoring

I went to DC yesterday to meet the winners of the first stage of the  Hewlett Foundation’s Automated Student Assessment Prize. It was a very interesting experience. For this stage, the task was to grade a range of essays that had been selected by the organizers, and for which human scores were available.

The first thing that struck me is that the winning teams were primarily data mining experts with an ability to pick up NLP and educational assessment research as needed. Some of them said they had read our papers, others that they had made everything up on their own. 

I’m enjoying reading Kaggle competitors’ experience reports and fitting the ideas into my thinking. One obvious practice is that teams often consist of several people, each of whom has a complete running system. The pattern is that competitors run separately for a while, then coalesce into teams who ensemble together their systems. If the systems are complementary, the collective does better than the individual components.

 

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

Categories

%d bloggers like this: