Our goal in PARLANCE is to design and build mobile applications that approach human performance in conversational interaction, specifically in terms of the interactional skills needed to do so. These skills will include recognising and generating conversational speech incrementally in real-time, adapting to new information domains without manual intervention, and reacting to new communicative situations. All of these skills will be learned or adapted using real data, and we will use them to build systems for Voice-enabled Interactive Search and Information Provision applications.
Current search engines, such as Google, work well only if the user has a single search goal and does not have multiple trade-offs to explore. For example, standard search works well if you want to know the phone number of a specific business but poorly if you are looking for a house with several different search criteria of varying importance, e.g. number of bedrooms versus bathrooms versus price etc.. The latter requires the user to collaborate conversationally over several turns.
PARLANCE differs fundamentally from previous approaches based on logic and symbolic reasoning, which are often extremely fragile in the face of high error rates and uncertain user behaviour. Instead, PARLANCE will represent all information used by the various components of the system probabilistically viadynamic Bayesian Networks, and will optimise decisions even in the presence of noise. In addition, PARLANCE will treat incremental dialogue processing (i.e. where dialogue events are recognised below the turn level) within a statistical framework for the first time. The PARLANCE prototype systems will be deployed in the three most widely spoken languages in the world (English, Spanish and Mandarin) in order to maximise the impact of the project.
P. Alexopoulos, M. Wallace (2013), “Towards Learning Personalized Semantic Relevance Paths In Dialogue Systems”, 8th International Workshop on Semantic Media and Adaptation, Bayonne, France, 2013.
Hastie H., M.A Aufaure, P. Alexopoulos, H. Cuayáhuitl, N. Dethlefs, M. Gasic, J. Henderson, O. Lemon, X. Liu, P. Mika, N. Ben Mustapha, V. Rieser, B. Thomson, P. Tsiakoulis, Y. Vanrompay, B. Villazon-Terrazas, S. Young (2013), “Demonstration of the Parlance system: a data-driven, incremental, spoken dialogue system for interactive search”, Demo at 14th Annual Meeting of the Special Interest Group on Discourse and Dialogue, Metz, France, 22-24 August 2013. [BibTex, PDF]
P. Alexopoulos, J.M. Gómez-Pérez (2013) “Interactive Acquisition of Fuzzy Ontological Knowledge in Dialogue Systems”. In Proceedings of the 7th international conference on Knowledge capture (K-CAP ’13). Banff, Canada. [BibTex, PDF]
P. Alexopoulos, Marie-Aude Aufaure, Nesrine Ben-Mustapha, Hugues Bouchard, James Henderson,, Beibei Hu,, Peter Mika,, Joel Lang, Jose Manuel Gomez-Perez, Yves Vanrompay (2013), “Semantic Hyperlocal Search for Parlance Mobile Spoken Dialogue System”, Poster presentation at the 10th Extended Semantic Web Conference, Montpelier, France, May 26th – 30th, 2013. [BibTex, PDF]