Classical Strategies and Concepts in Causal Discovery

on YouTube ↗01h 02m

" summary

Overview of causal discovery methods with concrete claims: PC algorithm uses a staged schedule of conditional independence tests (start with marginals, grow conditioning sets), removes edges, orients unshielded colliders and applies four acyclicity rules to output a CP-DAG (Markov equivalence class). Assumptions: causal sufficiency, acyclicity, faithfulness; issues: determinism or exact linear cancellations break faithfulness, finite-sample order-dependence matters. Complexity scales polynomially in p given bounded degree; PC is consistent with a shrinking alpha and correct tests.

AI-written summary · about

" chapters (8)

1. 0:00Introduction to the lecture
2. 1:20Overview of causal discovery
3. 4:23Methodological foundations
4. 10:44Markov property and methods
5. 15:22The faithfulness assumption
6. 22:15Colliders and equivalence
7. 31:23The PC algorithm
8. 48:53Algorithm steps and summary

" related talks

• 31mIntroduction to Causal Inference using pgmpyAnkur Ankan · PyData Amsterdam 2024
• 48mCausal Discovery in the Quantum ContextSally Shrapnel · Quantum Physics and Statistical Causal Models (Simons Institute workshop, 2022)
• 24mA Friendly Introduction To CausalityAlex Dimakis · MLconf SF 2016
• 29m🧠🧹 Causality - Mental Hygiene for Data ScienceEyal Kazin · PyData Global 2024
• 31mCollaborative Causal Discovery With Atomic InterventionsShiva Kasiviswanathan · Algorithmic Aspects of Causal Inference 2022
• 43mA Practical Guide to Causality in Python (For The Perplexed)Aleksander Molak · PyData NYC 2023
• 01h 01mThe Mathematics of Causal Inference, with Reflections on Machine Learning and the Logic of ScienceJudea Pearl · Symposium on Visions of the Theory of Computing 2013
• 31mLearning And Testing Causal Models: A Property Testing ViewpointArnab Bhattacharyya · Learning from Interventions 2022