Machine-learning breakthrough paves way for medical screening

I love seeing things like this - training machines to do repetitive tasks so that people can apply their time, effort and creativity elsewhere. After all, isn't that what computers are for?

University of Alberta computing science PhD graduate Siamak Ravanbakhsh published his research in the scientific journal PLOS ONE on an automated process that increases the speed and accuracy of producing a person’s metabolic profile from a sample of biofluid such as blood serum or cerebrospinal fluid using NMR spectrometry. Such a profile can provide a functional read-out of the developmental, physiological or pathological state of a biological system. This ability to quickly and accurately measure the levels of various molecules in a single sample of biofluid marks a significant advance for simple, cost-effective, predictive medical screening.