Paper: An Evaluation Of Strategies For Selective Utterance Verification For Spoken Natural Language Dialog

ACL ID A97-1008
Title An Evaluation Of Strategies For Selective Utterance Verification For Spoken Natural Language Dialog
Venue Applied Natural Language Processing Conference
Session Main Conference
Year 1997
Authors

As with human-human interaction, spoken human-computer dialog will contain situa- tions where there is miscommunication. In experimental trials consisting of eight dif- ferent users, 141 problem-solving dialogs, and 2840 user utterances, the Circuit Fix-It Shop natural language dialog system misin- terpreted 18.5% of user utterances. These miscommunications created various prob- lems for the dialog interaction, ranging from repetitive dialog to experimenter in- tervention to occasional failure of the dia- log. One natural strategy for reducing the impact of miscommunication is selective verification of the user's utterances. This paper reports on both context-independent and context-dependent strategies for utter- ance verification that show that the use of dialog context is crucial for in...