People often use instructional analogies to introduce new concepts or explanations. Examples of these analogies can be found in instructional science texts and guides for inservice teachers. Building software that can understand instructional analogies has two potential benefits. First, it would enable an individual to convey new or existing qualitative concepts to a machine using analogies, thereby allowing them to explain ideas that may be cumbersome to describe otherwise. Second, a computational model of interpreting multimodal instructional analogies can be used as the basis of intelligent educational software, given that the use of analogies to teach science is a common and recommended practice for teachers. In this paper, we describe work in progress on techniques using qualitative spatial representations for interpreting analogies that are composed of simplified English text and sketches, illustrated by a running example.
|Original language||English (US)|
|Number of pages||8|
|State||Published - 2015|