Causal Reasoning Tutorial
What is causality? How can we answer causal questions with machine learning, statistics, and data science? This tutorial will explore the answers to these questions. Topics include: causality as a hypothetical intervention; the causal hierarchy of observe, act, imagine; causal graphical models (and how they are different from Bayesian networks); backdoor adjustment and the backdoor criteria; structural causal models and counterfactuals; estimating counterfactuals with abduction; the potential outcomes framework (and its relationship to structural causal models). Throughout the tutorial we will discuss where ML and causality meet, highlighting ML algorithms for causal inference and clarifying the assumptions they require. The aim of the tutorial is to prepare researchers to dive deeper into ML and causality.
Fairness, Privacy, and Ethics in Data Science Tutorial
University of Pennsylvania