UNSILO and the Organisation for Economic Co-operation and Development (OECD) have worked together on a proof of concept designed to improve content discovery and collection-building from the hundreds of thousands of books and papers available on the OECD Publishing platform, using UNSILO’s artificial intelligence technology. The solution is delivered using UNSILO’s concept extraction engine, which uses semantic tools and machine learning to extract and analyze the key concepts from OECD’s publications. The UNSILO tool works in any subject domain, and the concepts extracted do not require a pre-existing taxonomy or classification system.
The first deliverable of the Proof of Concept is based on UNSILO Classify, a desktop application that can automatically create subsets of existing collections around any broad or narrow subject. Classify provides easy and quick means to create and maintain new and innovative content packages that can support communication and dissemination activities, without losing control of the process. The tool not only provides for the selection of content to be adjusted, but also enables the in-house subject matter experts to control the degree of automation of the subject extraction, thereby reducing the cost and time of manually creating collections of content.
“We are very pleased to have built with UNSILO this Proof of Concept”, commented Toby Green, Chief Operating Officer of the OECD Public Affairs and Communication Directorate. “We were impressed that UNSILO solved a very practical problem for us in demonstrating how we can re-classify all 250,000 content items at OECD to any topic in a flexible way adapted to user needs regardless of context”.
Follow the link below to download the full Press Release: