Explore causal fairness analysis in this comprehensive tutorial presented by Elias Bareinboim from Columbia University at FAccT 2021. Delve into Structural Causal Models (SCM) and their applications in real-world scenarios, including Berkeley admissions, COMPAS prediction, and UCI Adult dataset. Examine the Causal Fairness Framework, standard fairness models, and discrimination analysis. Investigate counterfactual effects, including direct, indirect, and spurious effects, through thought experiments. Gain insights into causal explanation formulas and their relevance in addressing fairness issues in machine learning and artificial intelligence.
Translation Tutorial - Causal Fairness Analysis
Association for Computing Machinery (ACM) via YouTube
Overview
Syllabus
Structural Causal Model (SCM)
SCM M + Causal Diagram G
Berkeley admission Students apply for university's admission (7), and choose specific departments to which they wish to
COMPAS prediction. Northpointe are trying to predict represents the age, variable represents prior convictions, and
UCI Adult). The US census data records whether a person
Causal Fairness Framework: Step 2
Causal Fairness Framework (Summary)
The "Standard Fairness Model"
Discrimination in UCI Adult
Counterfactual Direct Effect
Counterfactual Indirect Effect
Thought Experiment III
Counterfactual Spurious Effect
Causal Explanation Formula
Taught by
ACM FAccT Conference
