This site has been archived on
The Community Research and Development Information Service - CORDIS
Information & Communication Technologies

Language Technologies


back to overview

Project factsheets will no longer be updated.  All information relevant to the project can be found on the CORDIS factsheet .  This is updated on a regular basis with public deliverables, etc.

PARLANCE - Probabilistic Adaptive Real-Time Learning And Natural Conversational
Engine

287615 - STREP

parlance-logo.jpg

At a glance

FP7-ICT-2011-7 - Language technologies

Challenge

The project aims at designing and building mobile applications that approach human performance in conversational interaction for hyper-local search. The project will provide a statistical framework for modelling all aspects of a dialogue, including speech-in and -out, dialogue modelling, natural language understanding and generation and user modelling with dynamic ontologies. This framework will enable the system to adapt to different users with different goals and evolve without the need for hand-coded rules. The project also involves building a "micro-turn taking" architecture to allow for more fluid, natural interaction. The system will work monolingually in 3 languages: English, Mandarin and Spanish.

Objectives and Innovation

The voice-based web search service aims to be more natural in terms of conversational interaction compared to traditional search. Current search engines 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.

Changing local contexts including geographic location will inform the interaction, optimising strategies even when the information available contains noise. This approach is fundamentally different to previous approaches based on logic and symbolic reasoning, which are often extremely fragile in the face of high error rates and uncertain user behaviour.

The project aims to revolutionize the way people search for information and the way spoken dialogue systems are developed and maintained.

The target group of the project

The voice-based web search service is operated either on simple mobile phones or on smartphones and is meant to be used by the general public. It could be used by people looking for services or searching a specific location, in particular in the context of tourism.

The result Impact

The result of the project will be a system for interactive voice-based web search service that can be used by people on the move. PARLANCE will provide unique statistical processing technology, which allows:

- greater robustness to recognition errors and noise,
- automatic strategy optimisation,
- ability to continuously learn on-line.

Impact

The project should enable faster adoption of spoken dialogue systems into the marketplace, enabling the supply of better services to citizens and businesses across language barriers. It should also allow users to access web-based information with basic mobile phones, enabling access to vital information, such as natural disaster information, in developing countries where mobile phones are ubiquitous. The coverage of Mandarin will provide information tailored to Chinese tourists in Europe (54 million in 2010) thus boosting the economy.

Co-ordinator

Contact Person:

Name: Dr Helen Hastie
Tel: +44 131 451 3344
E-mail

Organisation: HERIOT-WATT UNIVERSITY

More»
 







 

 

 

 

back to overview



This page is maintained by: Susan Fraser (email removed)