RDFox: managing semi-structured data
The internet contains a huge amount of semi-structured data – data which cannot easily be sorted into tables with pre-defined categories, and is therefore difficult to manage. However, researchers at the University of Oxford’s Department of Computer Science have developed methods to manage and exploit this type of data.
RDFox is a semi-structured data management system developed at Oxford during a five-year EPSRC Early Career Fellowship awarded to Professor Boris Motik. Semi-structured data management is important because it allows data to be combined from disparate sources into meaningful and valuable information. Query answering is a key part of data management systems; an important goal of RDFox is to effectively use RAM to improve query answering times and to exploit certain types of hardware to speed up processing.
EPSRC IAA funding supported two secondments which explored the applications of RDFox in industry. These resulted in the spinning out of two companies, and a licence to RDFox being sold to a company.
In the first project, Professor Motik’s team collaborated with several companies to develop RDFox and demonstrate its capabilities. Notably, a project with Kaiser Permanente, a US healthcare provider (HMO), won the best in-use paper award at the International Semantic Web Conference (ISWC) 2016. HMOs in the US are required to deliver measurement results to the US authorities on their quality of care. Using readily available hardware, RDFox was able to extract results in roughly 30 minutes compared to the 105 minutes needed for Kaiser Permante’s existing solution, as well as discovering several discrepancies in the reports produced by their internal software. The project thus confirmed the great benefits of semantic technologies in the area of healthcare compliance.
This work led to the spin-out of Oxford Semantic Technologies, which aims to bring cutting-edge research in semantic web technologies (such as the methods used by RDFox) to industry. Kaiser Permanente and Oxford Semantic Technologies are currently discussing a follow-up project which aims to turn the system into a comprehensive solution for compliance in healthcare. Oxford Semantic Technologies has raised nearly £1m, and is currently building a prototype that can be shown to investors to raise the sum needed for serious development of RDFox.
A follow-up project sought to develop an approach for content selection, using RDFox, to be applied to personalised broadcasting. The work was done in collaboration with Sixth Stage, a media company. Professor Motik, Professor Bernardo Cuenca Grau and their team developed a proof-of-concept prototype to demonstrate the potential of RDFox for content selection. The prototype was shown to executives of the BBC, and the spin-out company Covatic resulted from this venture.
Covatic seeks to “revolutionise the audience experience by enabling effortless discovery of content”. The software aims to increase revenue for the user company through means such as a bespoke channel for each user, real-time targeted advertising and performance monitoring, and automated tagging and description of content that can change over time. The intended effect will be to grow the audience, and increase time spent on the company’s services. To date, Covatic has raised around £600,000 from investors.