Causal Inference Colloquium II

Why Causation Matters

Stanford second colloquium on machine learning and causal inference

Monday, April 29, 2019


SIEPR · 366 Galvez St.

Koret-Taube Conference Room

Machine learning methods, both as a foundation for AI and used to find meaningful patterns in “Big Data”, are transforming multiple sciences, with some claiming that they can replace the traditional scientific method. However, an attribute of their power and source of societal fascination – an ability to explore complex data without human input or causal knowledge– is also a source of problems, such as nontransportable predictions, perpetuation of societal inequities and unreliable estimates of intervention effects.

This colloquium will explore the methods and value of combining causal knowledge and methodology with machine-learning algorithms to generate reliable health and social knowledge. It will be organized into three sessions: Prediction, Heterogeneity & Causal estimation, and Design. Each session will have 3 speakers, with substantial time for audience discussion. The colloquium is intended to promote a robust cross-disciplinary dialogue with speakers coming from the fields of epidemiology, statistics, informatics, computer science, law, medicine, and attendees from yet more disciplines. We look forward to your attendance and participation.

Suggested reading: It is time to learn from patient like mine

First colloquium: Is prediction enough?


Registration: We have reached capacity. The colloquium will be recorded and the videos will be published on this site.

Parking: Please use the code '1146' at the Galvez lot (294 Galvez St. Stanford, CA) for $8 parking. Here is a map of the Galvez parking lot to the SIEPR building.

Keynote Speaker: Jennifer Hill, PhD