Skip to Main Content
Relevance has so far been a major evaluation criterion in information retrieval. Treating context of search results as first-class user information needs, we explore a new direction of search in the light of information comprehension, which is a new dimension of relevance criteria. In this paper, we argue that it is entailment relations that enable the modeling of media comprehension at the conceptual level. We also propose a method to construct an entailment ontology for e-learning and show its application to lecture material retrieval, which aims to identify, for user queries, not only precise search targets inside media streams but also their surrounding context for enhancing comprehension (of the search targets).