S. Keizer

Dialogue act recognition under uncertainty using bayesian networks

In this paper we discuss the task of dialogue act recognition as a part of interpreting user utterances in context. To deal with the uncertainty that is inherent in natural language processing in general and dialogue act recognition in particular we use machine learning techniques to train classifiers from corpus data. These classifiers make use of both lexical features of the (Dutch) keyboard-typed utterances in the corpus used, and context features in the form of dialogue acts of previous utterances. In particular, we consider probabilistic models in the form of Bayesian networks to be proposed as a more general framework for dealing with uncertainty in the dialogue modelling process.

Multidimensional Dialogue Management

In this paper we present an approach to dialogue management that supports the generation of multifunctional utterances. It is based on the multidimensional dialogue act taxonomy and associated context model as developed in Dynamic Interpretation Theory (DIT). The multidimensional organisation of the taxonomy reflects that there are various aspects that dialogue participants have to deal with simultaneously during a dialogue. Besides performing some underlying task, a participant also has to pay attention to various aspects of the communication process itself, including social conventions. Therefore, a multi-agent approach is proposed, in which for each of the dimensions in the taxonomy a specialised dialogue act agent is designed, dedicated to the generation of dialogue acts from that particular dimension. These dialogue act agents operate in parallel on the information state of the system. For a simplified version of the taxonomy, a dialogue manager has been implemented and integrated into an interactive QA system.