ECAI 2004 Conference Paper

[PDF] [full paper] [prev] [tofc] [next]

A Speech Architecture for Personal Assistants Applied in a Knowledge Management Context

Emerson Cabrera Paraiso, Jean-Paul Barthes, Cesar Tacla

This paper describes the design and the problems encountered while developing a speech and natural language dialog interface for Personal Assistants. Personal Assistants are agents that help human users (often referenced as masters) to do their daily work. The particular skills of a Personal Assistant are devoted to understanding its master, and presenting the information intelligently and in a timely manner. We believe that the use of speech will facilitate the interaction between the user and his machine, since the user may speak with the agent using his own terms, increasing the quality of the assistance. The main difficulty however is to handle spoken natural language, understanding their actual context. To do so, we restrict the exchanges to Directives Speech Act classes: inform, request, or answer. In addition, the knowledge handling is crucial for the effectiveness of such interfaces. We are using a set of task and domain ontologies, separating out domain and task models for reasoning. In this work, we present such an architecture for a multi-agent system applied to knowledge management. As a clear result of the conversational speech interface, we could expect a reduction in the cognitive load of the user and an increase in the quality of assistance.

Keywords: Intelligent User Interfaces, Spoken Dialogue, Personal Assistants, Knowledge Management

Citation: Emerson Cabrera Paraiso, Jean-Paul Barthes, Cesar Tacla: A Speech Architecture for Personal Assistants Applied in a Knowledge Management Context. In R.López de Mántaras and L.Saitta (eds.): ECAI2004, Proceedings of the 16th European Conference on Artificial Intelligence, IOS Press, Amsterdam, 2004, pp.971-972.


[prev] [tofc] [next]


ECAI-2004 is organised by the European Coordinating Committee for Artificial Intelligence (ECCAI) and hosted by the Universitat Politècnica de València on behalf of Asociación Española de Inteligencia Artificial (AEPIA) and Associació Catalana d'Intel-ligència Artificial (ACIA).