19–21 Jun 2013
University of Geneva
Europe/Zurich timezone

Semantic Session: Detecting knowledge-level claims in research articles

20 Jun 2013, 10:00
30m
R380 (Uni Mail)

R380

Uni Mail

Speaker

Mrs Ágnes Sándor (Xerox Research Centre Europe)

Description

The research article genre has developed a specific structure for the effective communication of scientific results. The title, the abstract, more or less standardized section types, etc. are all structural elements in the service of facilitating comprehension and searchability. These elements are apparent for the reader through formatting, and and their markup makes it possible for search algorithms to take advantage of them in relevance ranking. In this presentation we propose another type of content element that can facilitate comprehension and search: knowledge-level claims. We define knowledge-level claims as discourse elements in articles that indicate the status of scientific propositions within the state of the art. Knowledge-level claims significantly contribute to comprehension, they are usually rhetorically salient, but traditionally they are not made prominent through formatting or markup. We show some applications where the automatic detection of knowledge-level claims has been used for enhancing both comprehension and search, and we indicate some further potential applications.

Presentation materials