Net benefit, calibration, threshold selection, and training objectives for algorithmic fairness in healthcare

1 code implementation • 3 Feb 2022 • Stephen R. Pfohl, Yizhe Xu, Agata Foryciarz, Nikolaos Ignatiadis, Julian Genkins, Nigam H. Shah

A growing body of work uses the paradigm of algorithmic fairness to frame the development of techniques to anticipate and proactively mitigate the introduction or exacerbation of health inequities that may follow from the use of model-guided decision-making.

Decision Making Fairness

Empirical Bayes mean estimation with nonparametric errors via order statistic regression on replicated data

2 code implementations • 14 Nov 2019 • Nikolaos Ignatiadis, Sujayam Saha, Dennis L. Sun, Omkar Muralidharan

We study empirical Bayes estimation of the effect sizes of $N$ units from $K$ noisy observations on each unit.

Methodology

Covariate-Powered Empirical Bayes Estimation

4 code implementations • NeurIPS 2019 • Nikolaos Ignatiadis, Stefan Wager

We study methods for simultaneous analysis of many noisy experiments in the presence of rich covariate information.

Confidence Intervals for Nonparametric Empirical Bayes Analysis

1 code implementation • 7 Feb 2019 • Nikolaos Ignatiadis, Stefan Wager

In an empirical Bayes analysis, we use data from repeated sampling to imitate inferences made by an oracle Bayesian with extensive knowledge of the data-generating distribution.

Methodology

Covariate powered cross-weighted multiple testing

3 code implementations • 18 Jan 2017 • Nikolaos Ignatiadis, Wolfgang Huber

Here, we show how to use information potentially available in the covariates about heterogeneities among hypotheses to increase power compared to conventional procedures that only use the $P_i$.

Methodology