Exploiting User Expertise in Answer Expression

Previous natural language help systems have not taken into account the user’s knowledge when formulating answers. Such pragmatic information is needed to formulate more concise and helpful answers. By not repeating things that the user already knows, a system can provide more succinct answers that, because they focus on pertinent new facts, are easier to understand. A users’s prior knowledge also allows a system to utilize special teaching formats such as similes. This process of refining answers using pragmatic information is called answer expression. It has been implemented in the UCExpress component of UC (UNIX Consultant), a natural language system that helps users solve problems in using UNIX. UCExpress separates answer expression into two phases: pruning and formatting. During pruning, subconcepts of the answer are marked as not needing generation when they are already known by the user, or marked as candidates for generating anaphora or ellipsis when they are part of the conversational context. During formatting, UCExpress uses information about the user’s prior domain knowledge to select among specialized expository formats, such as similes and examples, for expressing information to the user. These formats allow UCExpress to present different types of information clearly and concisely.