Machine Learning and Crowdsourcing Made Easy for Physicists

Machine Learning and Crowdsourcing Made Easy for Physicists

Qi Feng (McGill U.)

May 4, 2017 19:00

Everybody Can Teach a Machine a Little Physics

Abstract: The interest in machine learning began in the 1950s, a few years after computers were invented, but only recently has significant progress in its research and application been made. Such progress is made possible by low-cost computation and large datasets, the latter of which often contains contribution from humans. The popularity of machine learning and the hard work of programmers led to many free softwares that can be easily used to solve a wide spectrum of problems, including some physics ones.

I will give a basic introduction to machine learning and show examples of its application in physics. Some of these physics problems are tackled via a citizen-science/crowdsourcing approach, where science enthusiasts can inspect physics data (often in the form of images) and provide feedback that is used to train machine learning models. An excellent platform called Zooniverse ( hosts many citizen-science projects. Click the link, and you can start to learn about a certain scientific topic while some machines can learn from you!