Morning Session: Causality
Talkat Wrong at the Root: Racial Bias and the Tension Between Numbers and Words in Non-Internet Data · 2019-06-05..06-07·
summary
Three linked analyses: a labor market model shows Bayesian hiring on noisy signals plus group reputations creates a self-fulfilling disparity but introducing a temporary labor market with enforced statistical parity can (under technical conditions) produce a unique group‑symmetric steady state though timing and robustness are open questions; a causal modeling critique argues race/gender require a stable enough ontology because mislabeling mediators versus confounders breaks inference; clinical examples show the CKD‑EPI race correction raises estimated GFR for African Americans and alters transplant priority.
AI-written summary · about
chapters (7)
- 0:00Modeling labor discrimination
- 10:01Defining model boundaries
- 14:03Dynamic game interventions
- 22:40Causal models and categories
- 35:06Ontologies of race and bias
- 47:38Medical ethics and bias
- 54:26Statistical discrimination
related talks
34mInterpretability and Algorithmic Fairness
24mA Friendly Introduction To Causality
11mCausal Inference For Socio-Economic & Engineering Systems
29m🧠🧹 Causality - Mental Hygiene for Data Science
01h 01mThe Mathematics of Causal Inference, with Reflections on Machine Learning and the Logic of Science
31mA Gentle Introduction To Causal Inference
28mBeyond Prediction Performance: How Modeling Decisions Shape Fairness Outcomes in Statistical...
41mKeynote: Personalization, Bandits, and Causal inference