Can machines understand the scientific literature?

Every 15 seconds a new Scientific/Technical/Medical (STM) article is published – text, images, tables, diagrams and it’s impossible for anyone to keep up. We need machines to help and I’ll describe systems that can “understand” chemistry, evolutionary trees, etc. It’s much easier when everything is Open, and we are downloading and analysing papers in bioscience, astrophysics, clinical trials. I believe that Wikidata will become the primary means to index STM material and we can use this to build specialist search tools. Technically it’s becoming easy to create and deploy “text and data mining” (TDM) – or more widely “Content Mining” and very accessible to students (our youngest developer is 15 years old). But TDM has caused huge controversy in Europe. The UK made a small step in 2014 – it’s legitimate for personal non-commercial research – but similar legislation in Europe hit serious pushback from publishers. I’ll contend that Science is being held back by copyright. The audience is invited to participate so bring your laptops/mobiles and hopefully we’ll try some simple experiments.