If computer-based instructional systems are to reap the benefits of natural language interaction,they must be endowed with the properties that make human natural language interaction so effective. In this paper, I describe several features that distinguish human tutorial explanations from those produced by today's computer-based tutors. I argue that in order to build systems capable of structuring explanations naturally and appropriately interpreting and generating context-sensitive utterances, it is necessary to model and reason about the plans the speakers in a discourse are attempting to carry out. I then present a plan-based approach to natural language generation, and provide examples from several implemented systems that illustrate how the approach enables systems to participate effectively in natural language dialogues with their users.
|Number of pages||34|
|Journal||Journal of Artificial Intelligence in Education|
|Publication status||Published - 1 Nov 1996|