Deep Learning generates new insights from existing studies

By combining information from existing studies, new hypotheses and findings can be derived. In an ideal world, however, huge volumes of texts must first be analyzed and then cross-referenced for this to work effectively.

A research group from the Zurich University of Applied Sciences (ZHAW) is now seeking to make this process more straightforward. To this end, the team is making use of Deep Learning approaches and artificial neuronal networks, as detailed in an article included in “Transfer”, a magazine produced by ZHAW. Thanks to these methods, hypotheses can be derived from existing texts in an automated fashion.

In their project, the researchers are focusing on studies from the world of biomedicine. Their aim is, firstly, to discover more about natural drugs, and secondly, to generate an overall improvement in the Literature Based Discovery (LBD) method. The project is being funded by Health@N.