AKTors.org
AKTors.org AKTTechnologiesPublicationsRelated ProjectsPeople
AKTors.org AcquisitionModellingRetrievalReusePublishingMaintenance
Aqualog from The Open University

AquaLog / PowerAqua: AquaLog is a Natural Language Question Answering System for corporate semantics intranets. The next generation of AquaLog, PowerAqua, will be a QA system able to provide answers drawn from multiple heterogeneous ontologies on the Semantic Web in the large.  

AquaLog / PowerAqua fact-file

Owner  :  The Knowledge Media Institute, Open University
Researchers
(listed alphabetically)
 :  Vanessa Lopez, Enrico Motta, Victoria Uren, Marta Sabou
Description  :  http://kmi.open.ac.uk/technologies/aqualog/
Demonstration  :  http://aqualog.open.ac.uk:8080/aqualog/index.html
Builds on  :  WebOnto, MySQL, Sesame, Jena, Lucene, Gate, Apache
Used by  :  KMi Semantic Portal
Addresses challenges  :  Knowledge Retrieval

What's the Problem?

AquaLog uses ontologies as a powerful tool for factual Question Answering, where users are able to pose questions in their own terms, without having to worry about the structure of the ontology or learn a special query language.

Although portable it only uses one ontology at a time. However, there is a need to dynamically provide answers derived from many heterogeneous ontologies that can refer to different domains. PowerAqua belongs to a new generation of SW tools as it tries to answer questions asked in Natural Language by leveraging and truly exploit the large, distributed and heterogeneous Semantic Web on the fly, which implies that  traditional methods in ontology selection, mapping and semantic similarity measures need to be further advanced to balance the heterogeneity and large scale semantic data with run time performance. 

Towards a Solution

AquaLog takes a user's question in NL a turns it into a triple format. Reasoning about the ontology (taxonomy and relationships) and using linguistic resources, such as as Gate and WordNet, and information it has learned about the users through a learning mechanism, matches this triple against the ontology to return an answer drawn form semantic mark-up compliant with the input ontology.

Additionally, PowerAqua needs addresses issues that include: (!) Automatic selection of relevant ontologies and KBs to answer a query on the fly; (2) Check semantically sound mappings over the SW distributed nature; (3) Combine partial answers from different ontologies when needed. PowerAqua is based in a knowledge-based mapping algorithm call PowerMap.

AquaLog is distributed under the Gnu General Public License, and is available from Source forge. The first prototype of PowerAqua will be available in a few months time.

Semantic representation

View in the KMi Semantic Portal