NBIA Colloquium: Niels Richard Hansen
Title: Correlation is not causation; but what is causation then?
Abstract: Any decent intro statistics course drills the saying “correlation is not causation” into the heads of the students. They should learn that we cannot draw conclusions about causal relationships from associations in observed data only. But beyond this saying, the concept of causation was for most parts of the 20th century ignored, or even frowned upon, in the statistics community. We left students and all others with the impression that statistics was not helpful in answering questions about causality. In this talk I will cover the more recent explosion in the research in statistics and machine learning on formal causal theories and how we can learn about causality from data after all. I will illustrate the development with problems and results from my own research on temporal causal models.
Brief bio-sketch: Niels Richard Hansen is a Professor of Computational Statistics at Department of Mathematical Sciences, University of Copenhagen. His main interest is in methodology and algorithms for causal learning. He is a co-founder of Copenhagen Causality Lab that does research in the broader questions of causality at the intersection of AI
and statistics.
Refreshments in the NBIA Lounge after the talk!