Causal Machine Learning
This module covers causal inference and its role in machine learning, from standard approaches such as regression and propensity scores through to modern machine learning approaches for estimating heterogeneous treatment effects. During their private study, students should read the course literature, work on practical exercises, and write a project report for the examination.
Delivery
This module consists of 2-hour lectures per day for 5 days, plus a 1-hour tutorial per day.
Study Hours
Approximately 100 hours of study, of which 15 are contact hours and 85 are private study.
Software Used
Stata, Python
Credits
10 CATS credits
Programme Faculty
Programme Faculty
Key Outcomes
- —Ability to comprehend concepts and principles of causal inference.
- —Proficiency in deploying advanced machine learning methods for causal analysis.
- —Skills to perform rigorous causal analysis using machine learning techniques.
- —Hands-on experience in deploying advanced ML methods for causal analysis.
Desired Skills
- —Understand the basics of causal inference and its role in machine learning
- —Identify and estimate causal effects using common methods and algorithms
- —Apply causal machine learning to real-world problems in areas such as health, finance, and social sciences
- —Communicate effectively about causal inference and its applications in machine learning
Lecture Programme
- WK 01
Lecture 1: Introduction
- ·Simpson's Paradox
- ·Directed Acyclic Graphs (DAG)
- ·Potential Outcomes Causal Model
- ·AB-testing and RCT Models
- WK 02
Lecture 2: Standard approaches
- ·Regression and Double Orthogonalization
- ·Propensity Scores
- ·Doubly Robust Estimator
- WK 03
Lecture 3: Machine Learning approaches
- ·High-dimension and Variable Selection
- ·Post-Lasso and Post-Double-Lasso
- ·Double/debiased Machine Learning
- WK 04
Lecture 4: Heterogeneous Treatment Effects
- ·Meta-Learners
- ·Causal Forest
- ·Generic Machine Learning
- WK 05
Lecture 5: Methods
- ·Instrumental Variables
- ·Difference-in-Differences
- ·Synthetic Control
Assessment
Online coursework submission with the support of your module lecturer and tutor. All sessions are conducted live online at a scheduled time and recorded for review.