Grigory Chernov (HSE University, Laboratory of Experimental and Behavioral Economics; Tübingen University)
The talk introduces causal structure discovery from data: intuition behind graphical probabilistic models, the trade-off between the number of assumptions and the robustness of conclusions, basic approaches (constraint-based: PC, FCI; score-based: GES; semi-parametric: LiNGAM) and libraries Causal Discovery Toolbox and causal-learn. Plan: place of methods among other ways to work with data; why identification is needed; probabilistic intuition; short overview of methods.
Continue reading — it's free
Google or a one-time email code. No password.