Language Generation for Spoken Dialogue Systems [Invited Talk]

The goal of spoken dialogue systems (SDS) is to offer efficient and natural access to applications and services. A common task for SDS is to help users select a suitable option (e.g., flight, hotel, restaurant) from the set of options available. When the number of options is small, they can simply be presented sequentially. However, as the number of options increases, the system must have strategies for summarizing the options to enable the user to browse the option space. In this talk, we evaluate two recent approaches to information presentation in SDS: (1) the Refiner approach (Polifroni et al., 2003) which generates summaries by clustering the options to maximize coverage of the domain, and (2) the user-model based summarize and refine (UMSR) approach (Demberg and Moore, 2006) which clusters options to maximize utility with respect to a user model, and uses linguistic devices (e.g., discourse cues, adverbials) to highlight the trade-offs among the presented items.