Causal Machine Learning

  • Integrates causal inference principles with machine learning models to estimate treatment effects or counterfactual outcomes.
  • These ML-based methods are built for high-dimensional or flexible modeling but are tailored for questions of cause-effect rather than pure prediction.
  • They rely on combining outcome modeling with design-based techniques.
  • Often use sample splitting, meta-learners, and regularization to improve robustness and flexibility.

Overlap: Strong synergy with Adjusted Methods (e.g., doubly robust learning), Causal Discovery (for model structure), and Alternative Paradigms

Key feature: Learns causal effects using ML models

Usage: For flexible and scalable causal inference in high-dimensional settings


Table of contents