An easy method to make dialogue systems incremental

Abstract

Incrementality as a way of managing the interactions between a dialogue system and its users has been shown to have concrete advantages over the traditional turn-taking frame. Incremental systems are more reactive, more human-like, offer a better user experience and allow the user to correct errors faster, hence avoiding desynchronisations. Several incremental models have been proposed, however, their core underlying architecture is different from the classical dialogue systems. As a result, they have to be implemented from scratch. In this paper, we propose a method to transform traditional dialogue systems into incremental ones. A new module, called the Scheduler is inserted between the client and the service so that from the client's point of view, the system behaves incrementally, even though the service does not.

Optimising Turn-Taking Strategies With Reinforcement Learning

“…In our work, the turn-taking task is separated from the common dialogue management one and it is handled by a separated module called the Scheduler (Khouzaimi et al, 2014). A considerable asset of this architecture is that it can just be added to the agenda service in order to make it incremental.…”

“…A non-incremental dialogue system can be transformed into an incremental one by adding an extra module: the Scheduler (Khouzaimi et al, 2014). Its objective is to make turn-taking decisions (whether to take the floor or not).…”