Title :
Entrainment in Slovak collaborative dialogues
Author :
Benus, Stefan ; Levitan, Rivka ; Hirschberg, Julia ; Gravano, Agustin ; Darjaa, Sakhia
Author_Institution :
Inst. of Inf., Univ. in Nitra, Bratislava, Slovakia
Abstract :
Entrainment is a natural inclination of people who interact with each other to develop similar, matching, or synchronized forms of behavior. In spoken interactions, this was observed at many linguistic (syntactic structures, referring expressions, dialectal features) and para-linguistic (intensity, speech rate) levels. Numerous studies have shown that the degree of spoken entrainment correlates with several social conversational attributes such as task success, likability, attractiveness, power distribution, and others. This feature of human-human interactions promises great potential for improving the naturalness and social acceptability of voice applications in human-machine interactions. However, despite many findings of various forms of entrainment in various languages and cultures, there is very little work on analyzing the effect of language or culture on speech entrainment in comparable spoken corpora. This is a significant gap since many applications utilizing spoken interaction and entrainment will be deployed to be used by native and non-native speakers of many cultures and languages. A welcome exception filling this gap is a recent paper by Xia et al. (2014) comparing entrainment in Standard American English and Mandarin Chinese. In this paper we provide a first analysis of speech entrainment in Slovak based on comparable corpus to the ones analyzed in Xia et al. This work contributes to our understanding and future modeling of cross-cultural aspects of behavioral entrainment.
Keywords :
computational linguistics; cultural aspects; human computer interaction; interactive systems; natural language processing; speech processing; Slovak collaborative dialogue; culture; human-human interactions; human-machine interactions; naturalness improvement; non-native speakers; para-linguistic level; social acceptability improvement; social conversational attributes; speech entrainment; spoken entrainment; spoken interaction; spoken interactions; voice applications; Convergence; Feature extraction; Pragmatics; Robustness; Speech; Speech recognition; Standards; Slovak; entrainment; human-computer interaction; human-human dialogues; prosody;
Conference_Titel :
Cognitive Infocommunications (CogInfoCom), 2014 5th IEEE Conference on
Conference_Location :
Vietri sul Mare
DOI :
10.1109/CogInfoCom.2014.7020468