Corpus ID: 75455

On Fairness and Calibration

@inproceedings{Pleiss2017OnFA,
  title={On Fairness and Calibration},
  author={Geoff Pleiss and M. Raghavan and Felix Wu and J. Kleinberg and Kilian Q. Weinberger},
  booktitle={NIPS},
  year={2017}
}
  • Geoff Pleiss, M. Raghavan, +2 authors Kilian Q. Weinberger
  • Published in NIPS 2017
  • Computer Science, Mathematics
  • The machine learning community has become increasingly concerned with the potential for bias and discrimination in predictive models. This has motivated a growing line of work on what it means for a classification procedure to be "fair." In this paper, we investigate the tension between minimizing error disparity across different population groups while maintaining calibrated probability estimates. We show that calibration is compatible only with a single error constraint (i.e. equal false… CONTINUE READING
    250 Citations

    Paper Mentions

    On Fairness in Budget-Constrained Decision Making
    • 3
    • Highly Influenced
    • PDF
    Statistical Equity: A Fairness Classification Objective
    • 2
    • PDF
    Leveraging Labeled and Unlabeled Data for Consistent Fair Binary Classification
    • 10
    • PDF
    Robust Fairness under Covariate Shift
    • PDF
    Recovering from Biased Data: Can Fairness Constraints Improve Accuracy?
    • 12
    • PDF
    A General Framework for Fair Regression
    • 7
    • PDF
    Group calibration is a byproduct of unconstrained learning
    • 1
    • PDF
    A Distributionally Robust Approach to Fair Classification
    • 2
    • PDF
    Eliminating Latent Discrimination: Train Then Mask
    • 2
    • PDF
    Taking Advantage of Multitask Learning for Fair Classification
    • 22
    • PDF

    References

    SHOWING 1-10 OF 40 REFERENCES
    On the relation between accuracy and fairness in binary classification
    • 72
    • PDF
    Equality of Opportunity in Supervised Learning
    • 1,192
    • Highly Influential
    • PDF
    An algorithm for removing sensitive information: application to race-independent recidivism prediction
    • 74
    • PDF
    Classifying without discriminating
    • F. Kamiran, T. Calders
    • Computer Science
    • 2009 2nd International Conference on Computer, Control and Communication
    • 2009
    • 214
    • PDF
    Learning Fair Classifiers
    • 33
    • PDF
    Learning Fair Representations
    • 716
    • PDF
    Inherent Trade-Offs in the Fair Determination of Risk Scores
    • 662
    • PDF
    Counterfactual Fairness
    • 419
    • PDF