Position: Research Associate
Office: Kilburn Building Room 2.116
Phone: +44 (0) 161 275 6137
Email: rshearer@cs.man.ac.uk
Personal website: http://v.cx
|
Position: Research Associate Office: Kilburn Building Room 2.116 Phone: +44 (0) 161 275 6137 Email: rshearer@cs.man.ac.uk Personal website: http://v.cx |
I am currently a Research Associate at the University of Manchester and am working towards a PhD in knowledge representation and automated reasoning, with a focus on reasoning systems for expressive description logics which scale efficiently to large ontology sizes. I am also a member of Manchester’s Information Management Group and that group’s Description Logic clique.
Before coming to Manchester I developed the Cerebra reasoner while with Network Inference (located in London when I joined, and Carlsbad, California when I left). That company has since been renamed to “Cerebra, Inc.” and acquired by WebMethods, which has itself been acquired by SoftwareAG; I have no idea what the status of the Cerebra reasoner is these days. During my employment I spent some amount of time disagreeing with many decisions of the W3C’s RDF Data Access Working Group over the design of SPARQL.
I also spent a few years with Transversal of Cambridge, England leveraging natural language processing for the purposes of web content management.
In my younger days I attended Brown University, where I founded and served as President for Technology House, the university’s first and only technology-oriented residential society.
My supervisors are Ian Horrocks and Uli Sattler. I share office 2.116 with Yevgeny Kazakov, Bernardo Cuenca Grau, Boris Motik, and Héctor Pérez Urbina. Other frequent collaborators include Birte Glimm, Charles Penwill, Bijan Parsia, and Matthew Horridge.
I currently manage the “Semantic Web Language Extensions” work package (number 2.5) of the EU KnowledgeWeb project, and in most cases serve as Manchester’s representative within other work packages.
I am also working with Boris Motik on a new hypertableau-based reasoning system called HermiT.
In support of continuous work with OWL ontologies and experimentation with new reasoning algorithms I developed Structured Ontology Format, an easily-parsed encoding for OWL ontologies which also provides a data model for direct use as an ontology API. This format includes an extended version of Manchester OWL Syntax.
I have also been involved with the Instance Store project, which attempts to apply standard tableau reasoning to very large ABoxes. The original Instance Store was developed by Daniele Turi (based on work with Sean Bechhofer and Ian Horrocks) but addressed only ABoxes without relations between individuals. Daniele, Lei Li, and Uli Sattler then did some work on “Instance Store 2” (iS2), which extends the approach to ABoxes with relations, in the hope that the potential efficiency problems which could result would be rare in practice or might be optimized away. I extended and optimized the initial iS2 implementation and also did some preliminary work on a rewrite which avoided the large cost associated with using an external database for all storage, but the project never extended beyond core reasoning components; practical query interfaces and robust libraries are not available. Dmitry Tsarkov is currently doing some work on repurposing the core logic of iS2 for computing precompletions of biological ontologies. I consider the HermiT approach to be much more easily amenable to reasoning for large ABoxes; hopefully that implementation will obsolete Instance Store completely in the near future.