Social talk capabilities for dialogue systems

Abstract

Small talk capabilities are an important but very challenging extension to dialogue systems. Small talk (or “social talk”) refers to a kind of conversation, which does not focus on the exchange of information, but on the negotiation of social roles and situations. The goal of this thesis is to provide knowledge, processes and structures that can be used by dialogue systems to satisfactorily participate in social conversations. For this purpose the thesis presents research in the areas of natural-language understanding, dialogue management and error handling. Nine new models of social talk based on a data analysis of small talk conversations are described. The functionally-motivated and content-abstract models can be used for small talk conversations on various topics. The basic elements of the models consist of dialogue acts for social talk newly developed on basis of social science theory. The thesis also presents some conversation strategies for the treatment of so-called “out-of-domain” (OoD) utterances that can be used to avoid errors in the input understanding of dialogue systems. Additionally, the thesis describes a new extension to dialogue management that flexibly manages interwoven dialogue threads. The small talk models as well as the strategies for handling OoD utterances are encoded as computational dialogue threads.