Cambridge Semantic Web Monthly Meetup

**********************************

Agenda: 2 Talks from New Brunswick, Canada

Talk 1: Valet SADI: Provisioning SADI Web Services for Semantic Querying of Relational Databases

Talk 2: WSReasoner: Strategically Using Faster Reasoners For Harder Logics

**********************************

Valet SADI: Provisioning SADI Web Services for Semantic Querying of Relational Databases

Abstract:
Semantic Querying (SQ) is emerging as an attractive approach for retrieval of data from relational and other conceptually similar databases, targeting users with limited or no technical expertise. Using SQ queries can be formulated using terminologies from a specific domain, which are then either translated in real time into the equivalent SQL queries, or executed against a materialised semantic database obtained by transforming the source relational data. This approach is suitable for non-technical users who are familiar with describing a domain using the terminologies in an ontology but lack expertise in writing SQL queries over complex relational schemas. As an alternative to the existing methods, we implement the vision of SQ on relational databases by deploying Semantic Web services over one or more databases and querying them with SPARQL queries. The approach based on manual Semantic Web service writing under-utilises easily manageable declarative mappings between source data schemas and domain ontologies. In this paper we introduce the Valet SADI framework to automate the creation of SADI Semantic Web services from declarative mappings, and demonstrate that Valet SADI, together with SADI query engines, establishes semantic querying as a viable, economical and user friendly way of querying relational databases. The running use case is the provision of a services on top of a hospital data warehouse to support surveillance queries for detection of hospital acquired infections.

ABOUT THE SPEAKER:

https://a248.e.akamai.net/secure.meetupstatic.com/photos/event/3/1/a/d/600_453192717.jpeg

Chris Baker is co-founder and CEO of IPSNP Computing Inc. established in 2011 to commercialize a unique query engine which provides fully integrated access to analytical software and online databases. Chris is also full professor and chair at the University of New Brunswick focused on knowledge management, data and service integration, semantic technologies, text mining, and web services. In addition to an academic career Chris held numerous positions in industry and government research labs. Up until 2008 he was the head of the Semantic Technology Group at the Data Mining Department, I2R, (A-STAR) Singapore and served as Chief Scientific Officer for Knorex Pte Ltd (Singapore). Prior to this he was a group leader of In-Silico Discovery at Ecopia BioSciences (now BELLUS Health) and Bioinformatics Manager at Concordia University in Montreal. He received postdoctoral training in micro and molecular biology at the University of Toronto and at Iogen Corporation in Ottawa. Chris has served as invited expert at the World Wide Web Consortium (W3C) for three years.

WSReasoner: Strategically Using Faster Reasoners For Harder Logics

WSReasoner reasons over OWL ontologies by applying a weakening and strengthening (WS) strategy. The strategy allows it to often use consequence based reasoning, which normally applies to ALCH, to reason over more expressive languages, including ALCHOI(D). When the strategy succeeds, efficient reasoners such as ConDOR can be applied. When it fails, it fails fast.

In this talk we review the WS strategy, focus on how it applies to nominals, and consider how this can be extended to number restrictions, which is a step on the way to the language SROIQ of OWL-2.

ABOUT THE SPEAKER

https://a248.e.akamai.net/secure.meetupstatic.com/photos/event/8/1/e/c/600_453213260.jpeg

Bruce Spencer’s research career spans 35 years in universities, government and private industry and he has been leader of projects that have gone on to successes including healthcare, social selling, and advanced training. He has published 75 peer-reviewed publications especially in automated reasoning, e-commerce, data mining service computing and data science. In 2013, the UNB entry into the OWL reasoning competition, developed by Weihong Song, Weichang Du and Bruce, won the "live" reasoning event.