Measuring Non-cooperation in Dialogue

Proceedings of COLING 2016, the 26th International Conference on Computational Linguistics: Technical Papers, pages 1925–1936, Osaka, Japan, December 11-17 2016. Measuring Non-cooperation in Dialogue Brian Pl¨uss Knowledge Media Institute The Open University Milton Keynes, UK brian.pluss@open.ac.uk Paul Piwek School of Computing and Communications The Open University Milton Keynes, UK paul.piwek@open.ac.uk Abstract This paper introduces a novel method for measuring non-cooperation in dialogue. The key idea is that linguistic non-cooperation can be measured in terms of the extent to which dia- logue participants deviate from conventions regarding the proper introduction and discharging of conversational obligations (e.g., the obligation to respond to a question). Previous work on non-cooperation has focused mainly on non-linguistic task-related non-cooperation or modelled non-cooperation in terms of special rules describing non-cooperative behaviours. In contrast, we start from rules for normal/correct dialogue behaviour – i.e., a dialogue game – which in principle can be derived from a corpus of cooperative dialogues, and provide a quantitative measure for the degree to which participants comply with these rules. We evaluated the model on a corpus of political interviews, with encouraging results. The model predicts accurately the degree of co- operation for one of the two dialogue game roles (interviewer) and also the relative cooperation for both roles (i.e., which interlocutor in the conversation was most cooperative). Being able to measure cooperation has applications in many areas from the analysis – manual, semi and fully automatic – of natural language interactions to human-like virtual personal assistants, tutoring agents, sophisticated dialogue systems, and role-playing virtual humans. 1 Introduction This paper describes a general method for measuring the degree of cooperation of dialogue participants’ behaviour. Central to the method is the idea, following Traum (1994) and Matheson et al. (2000), that in dialogue obligations are continually created and resolved. Our contribution is a proposal for measuring non-cooperation in terms of the degree to which dialogue participants deviate from the obligations that they acquire during the course of the dialogue. We focus on an application of the proposed method to political interviews in order to evaluate its validity. We developed this method to extend the state-of-the- art of computational dialogue modelling to cases in which the conversational flow is compromised to some extent but without reaching complete breakdown. Shedding light on the nature of linguistic non- cooperation in dialogue promises to yield a better understanding of conversation. The method can be used for the analysis – manual, semi and fully automatic – of natural language interactions and for applications such as human-like virtual personal assistants, tutoring agents, sophisticated dialogue systems, and role- playing virtual humans. In the remainder of this paper, we proceed as follows. In Section 2, we look at recent research in computational modelling of non-cooperative dialogue. We highlight the similarities and differences with the approach proposed in this paper. The next two sections then describe the two principal steps of our method. In Section 3, we introduce the first step. This step consists of segmentation of dialogue transcripts and coding of the …

Content truncated to 500 words