Causal Machine Learning

Description:
The course consists of two integrated parts:

• Part I (Lecture & Exercises): The first half of the semester introduces (i) the theoretical foundations of causal inference and the main strategies to identify causal effects, (ii) machine learning methods, and (iii) specific machine learning techniques for causal inference. Lectures are complemented by applied exercise sessions using Python.

• Part II (Seminar): In the second half of the semester, students independently work on a topic in causal machine learning. This includes presenting a research paper, implementing selected methods, and writing a seminar paper.

Content:

Foundations of Causal Inference
• Potential Outcomes Framework
• Randomised Controlled Trials (RCTs)
• Selection Bias and Identification
• Control Variable Strategies
• Instrumental Variables (IV)
• Difference-in-Differences (DiD)
• Regression Discontinuity Design (RDD)

Machine Learning Methods
• Prediction vs. Causation
• Regularised Regression (Lasso, Ridge)
• Model Selection and Overfitting
• Tree-Based Methods (Regression Trees, Random Forests)

Causal Machine Learning
• Post-Double Selection (PDS)
• Double / Debiased Machine Learning (DML)
• Orthogonalisation and Neyman Orthogonality
• Estimation of Average Treatment Effects (ATE)
• Heterogeneous Treatment Effects
• Simulation-based evaluation of estimators
• Advanced topics in Causal ML (deep learning, auto debiased ML, sensitivity analysis,
etc.)

For detailed description of the course, please see the syllabus. 

The PhD students may enroll in one or both parts of the course. Participation in the second part is open to students who have participated in the first part or who have acquired equivalent prior knowledge (in particular in advanced causal inference and machine learning).

Literature:
Please see the full literature list in the course syllabus.

Time & venue:
Lectures: Thursdays, 14.00 – 16.00, FU Berlin, Garystr. 35, Henry-Ford-Bau/ K II Konferenzraum
Exercise Sessions: Wednesdays, 16:00 – 18.00, Garystr. 35, Henry-Ford-Bau/ K II Konferenzraum
Seminar: Thursdays, 16:00 – 18.00, Garystr. 35, Henry-Ford-Bau/ K II Konferenzraum

Exam:
Seminar Paper, Presentation
Active Participation: Mandatory Test (pass/fail) after the lecture phase