Generating Effective Instructions: Knowing When to Stop

One aspect of Natural Language generation is describing entities so that they are distinguished from all other entities. Entities include objects, events, actions, and states. Much attention has been paid to objects and the generation of their referring expressions (descriptions meant to pick out or refer to an entity). However, a growing area of research is the automated generation of instruction manuals and an important part of generating instructions is distinguishing the actions that are to be carried out from other possible actions. One distinguishing feature is an action's termination, or when the performance of the action is to stop. My dissertation work focuses on generating action descriptions from action information using the SPUD generation algorithm developed here at Penn by Matthew Stone. In my work, I concentrate on the generation of expressions of termination information as part of action descriptions. The problems I address include how termination information is represented in action information and expressed in Natural Language, how to determine when an action description allows the reader to understand how to perform the action correctly, and how to generate the appropriate description of action information. Comments University of Pennsylvania Department of Computer and Information Science Technical Report No. MSCIS-98-35. This technical report is available at ScholarlyCommons: https://repository.upenn.edu/cis_reports/166 GENERATING EFFECTIVE INSTRUCTIONS: KNOWING WHEN TO STOP