tailieunhanh - Báo cáo khoa học: "REPAIRING REFERENCE IDENTIFICATION FAILURES BY RELAXATION"

The goal human-machlne of thls w o r k is the enrichment of mteractIons in a natural language envlronment. 1 W e w a n t to provide a f r a m e w o r k less restrictive than earlier ones by allowing a speaker leeway tn forming an utterance about a task and in determining the conversational vehicle to deliver it, A speaker a n d listener cannot be assured to have the s a m e beliefs, contexts, b a c k g r o u n d s or goals at each point in. | REPAIRING REFERENCE IDENTIFICATION FAILURES BY RELAXATION Bradley A. Goodman BBN Laboratories 10 Moulton Street Cambridge. Mass. 02238 ABSTRACT The goal of this work IS the enrichment of human-machine interactions in a natural language We want to provide a framework less restrictive than earlier ones by allowing a speaker leeway m forming an utterance about a task and in determining the conversational vehicle to deliver it A speaker and listener cannot be assured to have the same beliefs contexts backgrounds or goals at each point in a conversation. As a result difficulties and mistakes arise when a listener interprets a speaker s utterance These mistakes can lead to various kinds of misunderstandings between speaker and listener including reference failures or failure to understand the speaker s intention. We call these misunderstandings miscommunication Such mistakes constitute a kind of ill-formed input that can slow down and possibly break down communication Our goal IS to recognize and isolate such miscommunications and circumvent them This paper will highlight a particular class of miscommunication - reference problems - by describing a case study including techniques for avoiding failures of reference 1 Introduction Cohen. Perrault and Allen showed in their paper Beyond Question Answering 6 that . users of question-answering systems expect them to do more than just answer isolated questions they expect systems to engage m conversation In doing so. the system IS expected to allow users to be less than meticulously literal in conveying their intentions and it IS expected to make linguistic and pragmatic use of the previous discourse. Following in their footsteps we want to build robust natural language processing systems that can detect and recover from miscommunication. The development of such systems requires a study on how people communicate and how they recover from problems in communication. This paper summarizes the results of a .