Turn-taking phenomena in incremental dialogue systems

Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 1890–1895, Lisbon, Portugal, 17-21 September 2015. c⃝2015 Association for Computational Linguistics. Turn-taking phenomena in incremental dialogue systems Hatim Khouzaimi Orange Labs LIA-CERI, Univ. Avignon hatim.khouzaimi@orange.com Romain Laroche Orange Labs Issy-les-Moulineaux France romain.laroche@orange.com Fabrice Lef`evre LIA-CERI, Univ. Avignon Avignon France fabrice.lefevre@univ-avignon.fr Abstract In this paper, a turn-taking phenomenon taxonomy is introduced, organised accord- ing to the level of information conveyed. It is aimed to provide a better grasp of the behaviours used by humans while talking to each other, so that they can be method- ically replicated in spoken dialogue sys- tems. Five interesting phenomena have been implemented in a simulated environ- ment: the system barge-in with three vari- ants (resulting from either an unclear, an incoherent or a sufficient user message), the feedback and the user barge-in. The experiments reported in the paper illus- trate that how such phenomena are imple- mented is a delicate choice as their impact on the system’s performance is variable. 1 Introduction A spoken dialogue system is said to be incremen- tal when it does not wait until the end of the user’s utterance in order to process it (Dohsaka and Shi- mazu, 1997; Allen et al., 2001; Schlangen and Skantze, 2011). New audio information is cap- tured by an incremental Automatic Speech Recog- nition (ASR) at a certain frequency (Breslin et al., 2013) and at each new step, the partial available information is processed immediately. Therefore, the system is able to replicate a rich set of turn- taking phenomena (TTP) that are performed by human beings when talking to each other (Sacks et al., 1974; Clark, 1996). Replicating these TTP in dialogue systems can help to make them more efficient (e.g. (El Asri et al., 2014)) and enhance their ability to recover from misunderstandings (Skantze and Schlangen, 2009). Several contributions already explored different TTP like end-point detection (Raux and Eskenazi, 2008), backchannels (Meena et al., 2014; Visser et al., 2014), feedback (Skantze and Schlangen, 2009) or barge-in (Selfridge et al., 2013; Ghigi et al., 2014). However, these studies have been performed separately with no unified view and no comparison of respective merits, importance and co-influence of the different TTP. In order to have a better grasp on the concept of turn-taking in a di- alogue and a guideline for the implementation, we felt the need to introduce a taxonomy of these TTP. Our motivation is to clarify which TTP are inter- esting to implement given the task at hand. As an illustration, five TTP (which we assume have the best properties to improve the dialogue efficiency) have been implemented and compared in a slot- filling simulated environment. Section 2 introduces the TTP taxonomy and Section 3 describes the simulated environment, the experimental setup and the results. We then con- clude in Section 4. 2 Turn-taking phenomena taxonomy In linguistics and philosophy of language, a dis- tinction is made between two different levels of a speech act analysis: locutionary acts and illocu- …

Content truncated to 500 words