Attention, intentions, and the structure of discourse
Computational Linguistics
Artificial Intelligence
Unresolved issues in paragraph planning
Current research in natural language generation
Generating explanatory discourse
Current research in natural language generation
A problem for RST: the need for multi-level discourse analysis
Computational Linguistics
Planning text for advisory dialogues: capturing intentional and rhetorical information
Computational Linguistics
A pragmatics-based approach to ellipsis resolution
Computational Linguistics
A reactive approach to explanation
IJCAI'89 Proceedings of the 11th international joint conference on Artificial intelligence - Volume 2
| Hi-index | 0.00 |
In expository discourse, people sometimes ask questions that digress from the purpose of the discussion. A system that provides interactive explanations and advice must be able to distinguish pertinent questions from questions that digress. It must also be able to recognize questions that are incoherent. These types of questions require different treatment. Pertinent questions must be answered to achieve the discourse purpose. If the user asks a digressive question, the system may need to shift the focus of the discussion back to the original purpose. Incoherent questions signal a more serious misunderstanding that requires clarification and repair. The Interactive Discourse Planner (IDP) is designed to plan text to describe and/or justify a domain plan interactively. As a testbed, IDP plans text to discuss driving routes. IDP uses questions from the user to recognize how to extend its own text plan in a way that both satisfies its listener and achieves the system's discourse goal. In the process of recognizing ways to expand its own text plan, IDP can detect three types of digressions that the user can initiate with a question.