(16 Sep 2019) Cambridge University Press, one of the world’s largest and most prestigious academic publishers, and UNSILO, Denmark-based AI company, today signed an agreement by which UNSILO will use machine-learning tools to identify related content across around one million journal articles and book chapters from the vast Cambridge corpus.
UNSILO applies machine learning and AI tools to identify significant concepts from a text corpus.
These concepts form the basis of a wide range of solutions to publishing workflows, including
building subject collections, identifying related articles, finding relevant journals, and many other
areas. Uniquely, the UNSILO tools require no pre-existing taxonomy or ontology, and can be
applied to multiple subject collections. Using a process of continuous ingestion, new articles and
books are indexed and linked to other content on Cambridge Core within hours of publication.
Nik Louch, Platform Engineering Director at Cambridge University Press, commented:
When we designed and built Cambridge Core, the entirely new platform linking Cambridge books and journals, it was always our intention to bring market-leading and innovative tools to apply to our
content. The UNSILO related articles link, delivered from cloud servers via a simple API, is a good
example of how Core can be enhanced in this way. Because of the exceptionally wide subject range of our content, ranging from archaeology to neuroscience, and international law to Anglo-Saxon
studies, we needed a solution that did not require the lengthy construction of multiple related
ontologies and extensive manual tagging. Our trials in 2018 showed that users clearly appreciated
and made use of the content links provided.
Thomas Laursen, CEO of UNSILO, commented:
Since UNSILO’s foundation in 2014 we have worked with some of the largest names in publishing, but it is an honour to work with the longest-established academic publisher, Cambridge University Press, and its illustrious roster of authors. We are delighted we can be applying leading-edge
machine-learning tools to such a rich content collection, and we look forward to delivering more tools
via the Core platform in coming years.