On the use of expectations for detecting and repairing human-machine miscommunication

TL;DR

This paper explores how expectations in dialogue systems can detect and repair miscommunication by modeling future user utterances and pragmatic relations, improving system performance.

Contribution

It introduces a method using dialogic expectations to identify and address miscommunication in spoken language systems, demonstrated with the DIALOGOS system.

Findings

System performance improved with expectations-based detection

Miscommunication linked to pragmatic deviations and implicatures

Expectations aid in repairing dialogue errors

Abstract

In this paper I describe how miscommunication problems are dealt with in the spoken language system DIALOGOS. The dialogue module of the system exploits dialogic expectations in a twofold way: to model what future user utterance might be about (predictions), and to account how the user's next utterance may be related to previous ones in the ongoing interaction (pragmatic-based expectations). The analysis starts from the hypothesis that the occurrence of miscommunication is concomitant with two pragmatic phenomena: the deviation of the user from the expected behaviour and the generation of a conversational implicature. A preliminary evaluation of a large amount of interactions between subjects and DIALOGOS shows that the system performance is enhanced by the uses of both predictions and pragmatic-based expectations.