My research is concerned with the problem of turntaking in discourse, especially as applied to intelligent interfaces, such as advice-giving systems or software help systems. A limitation of many discourse systems is their need for explicit turn-ending signals (e.g. pressing a return key). In such systems, mid-turn interruptions are impossible, although there are practical examples of where mid...

A number of researchers have studied turn-taking offsets in human-human dialogues. However, that work collapses over a wide number of different turn-taking contexts. In this work, we delve into the turn-taking delays based on different contexts. We show that turn-taking behavior, both who tends to take the turn next, and the turn-taking delays, are dependent on the previous speech act type, the...

One intriguing feature of the human communication system is the interactional infrastructure it builds on. In both dyadic and multi-person interactions, conversation is highly structured and organized according to set principles (Sacks et al., 1974). Human adult interaction is characterized by a mechanism of exchange based on alternating (and relatively short) bursts of information. In the majo...

In this perspective article, we consider the relationship between experience sharing and turn-taking. There is much evidence suggesting that human social interaction is permeated by two temporal organizations: (1) the sequential framework of turn-taking and (2) the concurrent framework of emotional reciprocity. From this perspective, we introduce two alternative hypotheses about how the relatio...

We propose a simple model of turn-taking in an information state based approach to dialogue using TTR (Type Theory with Records). The information state (dialogue gameboard) contains an agenda formulated as a list of speech event types that the dialogue participant plans to realize. A novel aspect of the proposal is that the agenda also includes types of events that intuitively should be carried...

In this paper, reinforcement learning (RL) is used to learn an efficient turn-taking management model in a simulated slotfilling task with the objective of minimising the dialogue duration and maximising the completion task ratio. Turn-taking decisions are handled in a separate new module, the Scheduler. Unlike most dialogue systems, a dialogue turn is split into microturns and the Scheduler ma...