The British Medical Journal (www.bmj.com) has launched its first use of UNSILO technology. Among the many websites the BMJ provides, there are a number based around a subject – in this case, all about cancer (oncology). The site can be seen here.
The oncology portal lists some of the relevant BMJ journals, such as ESMO Open, and BMJ Learning, a collection of medical case reports.
Users interested in oncology topics have in the past had to search across all 50+ of the BMJ journals to find a relevant article. From today, the UNSILO service automatically identifies, using machine learning, any oncology-related articles from across the BMJ corpus. The 15 or so most recent articles are shown on the website, in a box marked “ recent articles”:
The website is automatically updated every 24 hours, which means a fresh set of articles appears each day.
Commented Isaac Jones, Digital Product Manager: “We started compiling lists of recent articles by hand, but trying to keep the lists up to date is very labour-intensive. In contrast, given that we have several hundred medical topics that our users are interested in keeping up to date with, this technology enables us to scale the entire process. We only require human input to define the concepts once, following which the system automatically identifies the updated articles. The result is a better service for our customers.”