About

Supervised machine learning necessitates reference sets composed of labeled data. For a given scientist, the addition of labels to complex, large-scale data is laborious, repetitive and error-prone due to size, composition and personal skew. The dedicated app SciSwipe helps scientists in the processing of complex data to simple images and distributes these images to the public. Each user can help train a computer by swiping and in effect labeling a small number of such images on a smartphone, dividing labor and increasing reliability. Crowdsourced exploration of complex, large-scale data holds great potential – a trained computer can mine nearly infinite amounts of data for clues to improve human health. SciSwipe: make it sexy and bring in the crowd.