Childhood pneumonia is the number one killer of children under the age of five worldwide. The disease is a particular challenge for those living in developing countries, where there is a lack of clinical expertise and appropriate equipment to diagnose the disease. Adopting a technique called machine learning, scientists at Oxford University have taken clinical data from children with pneumonia to 'teach' a machine to identify critical symptoms and diagnose future cases. This machine can be something as easily distributed as a mobile phone, giving those who don't have easy access to doctors... Like this podcast? Please help us by supporting the Naked Scientists