This episode explores causal inference through the lens of directed acyclic graphs (DAGs) and nonlinear structural equation models (SEMs). It highlights how these models provide a formal, nonparametric framework for understanding causal relationships, moving beyond simpler linear assumptions. The text introduces concepts like counterfactuals and conditional ignorability, explaining how they are derived from SEMs and verified using DAGs. It further details two graphical methods for identifying causal effects: the counterfactual DAG approach and Pearl's backdoor criterion, both aimed at finding adjustment sets to eliminate confounding. Finally, the authors discuss the implications of faithfulness assumptions in causal discovery, emphasizing the practical challenges of inferring causal structures from real-world data.DisclosureThe CausalML Book: Chernozhukov, V. & Hansen, C. & Kallus, N. & Spindler, M., & Syrgkanis, V. (2024): Applied Causal Inference Powered by ML and AI. CausalML-book.org; arXiv:2403.02467. Audio summary is generated by Google NotebookLM https://notebooklm.google/The episode art is generated by OpenAI ChatGPT

Ver todos los episodios