In machine learning, we use data to automatically find dependences in the world, with the goal of predicting future observations. Most machine learning methods build on statistics, but one can also try to go beyond this, assaying causal structures underlying statistical dependences. The hope is that this also allows prediction in certain situations where systems change, for instance by interventions.
The talk introduced the basic ideas of machine learning, and illustrated them with application examples. It argued that while machine learning and "big data" analysis currently mainly focuses on statistics; the causal point of view can provide additional insights.
Professor Bernhard Schölkopf is based at the Max Planck Institute for Intelligent Systems in Germany. He was given the 2014 Royal Society Milner Award in recognition of his pioneering work in machine learning which defined the field of “kernel machines”, now widely used in all areas of science and industry.
Enquiries: Contact the events team