Entity ranking track

The necessity to return entities instead of just documents or elements is becoming increasingly important. This track aims to provide a forum where researchers may compare and evaluate techniques for engines that return lists of entities.

The track will use the Wikipedia data, where it will exploit the category metadata on pages to construct entity sets. For example, consider categories "Dutch politicians" or "Art museums and galleries". The entities in such a set are assumed to correspond to those Wikipedia entries that are labelled with this category.

Three tasks are planned: (1) Entity Ranking, which aim is to return entities that satisfy a topic described in natural language text; (2) List Completion, where given a topic text and a number of examples, the aim is to complete this partial list of answers; and (3) Associative Ranking, which aim is to generate or complete a partial list of entities, that differs from a given list of entities, in one of its attributes.

Schedule

Organisers