The 2012 INEX Snippet and Tweet Contextualization Tasks

This paper reports on our current experiments involving the Snippet and Tweet Contextualization Tracks of the 2012 INEX competition. Most of this work in snippet generation extends our earlier (2011) approach, described in [4], which produced a top-ranked result. The source of the snippet in these experiments is the top-ranked focused element(s) of the document in question. Another approach is based on using the document itself as the source of the snippet. Having identified the source, the snippet is then generated based on simple basic methodologies described herein. We also describe our experiments in tweet contextualization, a new track for INEX in 2012.