Elsevier Life Science Solutions’ approach to machine-reading data is to help researchers bypass an article altogether using text-mining to read over four million full text papers, that recognise entities, and the relationship between them.
This data is extracted to create a database and a modelling interface that helps researchers with their work. An example he shares looks at how melanoma cells inhibit the local immune response. The first step is to find out from literature which proteins are secreted by the cells. The next step lies in figuring out which cells are impacted by these proteins. Then, the third step is to single out only the cells that are inhibited by these secreted proteins. Importantly, once the data is collated, the extracted sentences from the papers that indicate the result can be quickly assessed by a researcher, double-checking the relevancy or accuracy.