This item is available under a Creative Commons License for non-commercial use only
This paper presents ongoing research on convergence of speech features in human dialogues, in view of simulating this behaviour in spoken dialogue systems. The TAMA method (time-aligned moving average), previously used on monitoring convergence of acoustic prosodic (a/p) features, is applied to temporal properties of speech (between-turn pauses and overlaps). The results are compared to those of an older study on the same features.
Kousidis, S., Dorran, D.: Monitoring Convergence of Temporal Features in Spontaneous Dialogue Speech. 1st Young Researchers Workshop on Speech Technology University College Dublin. Dublin, Ireland, 2009.