Track

Causal ML

Goal: Learn how to estimate treatment effects from experiments and observational data, from potential outcomes and DAGs through propensity weighting, doubly robust and double ML estimators, instrumental variables, meta-learners, causal forests, and difference-in-differences.

Prereqs: ML Basics. Math for ML helps for identification arguments.

Status: done

Work through the steps in order. Bold links open YouTube.

Step	Concept	YouTube	Read
1	Correlation vs causation	1.3 - Correlation Does Not Imply Causation and Why	Wikipedia — Correlation does not imply causation
2	Potential outcomes	2.1 - What are Potential Outcomes?	Cunningham — Potential Outcomes and Randomization
3	Causal graphs	3.4 - Causal Graphs	Facure — Graphical Causal Models
4	Backdoor adjustment	4.6 - The Backdoor Adjustment	Pearl 2009 — Causal inference in statistics: An overview
5	Propensity scores and inverse probability weighting	6.4 - Propensity Scores and Inverse Probability Weighting (IPW)	Facure — Propensity Score
6	Doubly robust estimation	6.5 - Doubly Robust Methods, Matching, Double Machine Learning, and Causal Trees	Facure — Doubly Robust Estimation
7	Double / debiased machine learning	Double Machine Learning for Causal and Treatment Effects	Chernozhukov et al. 2016 — Double/Debiased Machine Learning for Treatment and Structural Parameters
8	Instrumental variables and LATE	8.4 - Nonparametric Identification of the Local ATE (LATE)	Imbens & Angrist 1994 — Identification and Estimation of Local Average Treatment Effects
9	Meta-learners and the X-learner	6.3 - TARNet and X-Learner	Künzel et al. 2019 — Metalearners for Estimating Heterogeneous Treatment Effects using Machine Learning
10	Causal forests	Conditional Average Treatment Effects: Forests	Wager & Athey 2018 — Estimation and Inference of Heterogeneous Treatment Effects using Random Forests
11	Difference-in-differences	9.2 - Difference-in-Differences Overview	Cunningham — Difference-in-Differences Fundamentals

Found a broken link or an unclear step? Report a problem with this track.