A shallow model of backchannel continuers in spoken dialogue
EACL '03 Proceedings of the tenth conference on European chapter of the Association for Computational Linguistics - Volume 1
Human-Computer Interaction
Computing backchannel distributions in multi-party conversations
EmbodiedNLP '07 Proceedings of the Workshop on Embodied Language Processing
A probabilistic multimodal approach for predicting listener backchannels
Autonomous Agents and Multi-Agent Systems
Generating nonverbal signals for a sensitive artificial listener
COST 2102'07 Proceedings of the 2007 COST action 2102 international conference on Verbal and nonverbal communication behaviours
Backchannel strategies for artificial listeners
IVA'10 Proceedings of the 10th international conference on Intelligent virtual agents
Hi-index | 0.00 |
In the long tradition of corpus based research on listener behavior, whether it entails linguistic analysis or social signal processing, many different tasks have been used during the recording of the corpus. So far in no study the task which has been given to the participants has been an independent variable and no studies have looked into the effect of this variable on listener responses. In this paper we present the results of our comparison between listening behavior elicited by procedural and narrative tasks which were used during the recording of our MultiLis corpus. We will show that listeners in the procedural tasks show more agreement in their responses than listeners in the narrative tasks. Furthermore we will show that the long procedural task elicits more responses per minute than the short procedural task. We will reflect on these results in light of cognitive load and grounding theory.