Equality of Opportunity in Supervised Learning.¶
Hardt, M., Price, & Srebro, N. (2016). Equality of Opportunity in Supervised Learning. Advances in Neural Information Processing Systems 29.
Cited by¶
4 citations across 4 artifacts.
Each citation links to the sentence it supports in the citing article.
Primes¶
- Equity
- Consider a credit-card fraud-detection algorithm that flags transactions as fraudulent based on learned patterns: unusual location, unusual amount, rapid successive transactions, etc., the kind of supervised classifier whose error structure Hardt, Price, and Srebro (2016) analyze under their equality-of-opportunity criterion.
This sourceFormalizes algorithmic impartiality as an equalized-odds constraint (predictor and protected attribute independent conditional on the outcome), shifting from removing the protected feature to constraining the function's dependence structure.
- Consider a credit-card fraud-detection algorithm that flags transactions as fraudulent based on learned patterns: unusual location, unusual amount, rapid successive transactions, etc., the kind of supervised classifier whose error structure Hardt, Price, and Srebro (2016) analyze under their equality-of-opportunity criterion.
- Fairness
- Machine learning & algorithmic fairness: Demographic parity (equal positive prediction rates across groups); equalized odds (equal true-positive and false-positive rates, formalized by Hardt, Price, and Srebro (2016)); calibration (prediction accuracy within groups); individual fairness (similar individuals treated similarly, in the sense of Dwork, Hardt, Pitassi, Reingold, and Zemel (2012)); counterfactual fairness (removing causal influence of protected attributes).
This sourceFormalizes equalized odds (equal true-positive and false-positive rates across protected groups) and the relaxed equal opportunity criterion, shifting fairness from 'remove the protected feature' to constraining the predictor's conditional dependence.
- Machine learning & algorithmic fairness: Demographic parity (equal positive prediction rates across groups); equalized odds (equal true-positive and false-positive rates, formalized by Hardt, Price, and Srebro (2016)); calibration (prediction accuracy within groups); individual fairness (similar individuals treated similarly, in the sense of Dwork, Hardt, Pitassi, Reingold, and Zemel (2012)); counterfactual fairness (removing causal influence of protected attributes).
- Impartiality
- Hardt, Price, and Srebro's (2016) equality-of-opportunity framework formalizes impartiality as an equalized-odds constraint on the conditional distribution of predictions across protected groups.
This sourceFormalizes algorithmic impartiality as an equalized-odds constraint on the conditional distribution of predictions across protected groups; shifts the framing from removing the protected feature (fairness-through-unawareness) to constraining the function's dependence structure.
- Hardt, Price, and Srebro's (2016) equality-of-opportunity framework formalizes impartiality as an equalized-odds constraint on the conditional distribution of predictions across protected groups.
Domain-specific¶
Verification¶
This reference passed the adversarial substantiation pipeline: it was checked to exist and to support the claim it is attached to. See how references were verified.
Registry ID ref:68c01c85a3d1 · see in the full table