Mitigation
holisticai.bias.mitigation is a python module mitigating bias in algorithms. Our classes cover pre-processing, post-processing and post-processing.
Pre-processing
|
Reweighing preprocessing weights the examples in each group-label combination to ensure fairness before classification. |
Learning fair representations finds a latent representation which encodes the data well while obfuscates information about protected attributes. |
|
CorrelationRemover applies a linear transformation to the non-sensitive feature columns in order to remove their correlation with the sensitive feature columns while retaining as much information as possible (as measured by the least-squares error). |
|
|
Variational Fair Clustering helps you to find clusters with specified proportions of different demographic groups pertaining to a sensitive attribute of the dataset (group_a and group_b) for any well-known clustering method such as K-means, K-median or Spectral clustering (Normalized cut). |
In-processing
|
Exponentiated gradient reduction is an in-processing technique that reduces fair classification to a sequence of cost-sensitive classification problems, returning a randomized classifier with the lowest empirical error subject to fair classification constraints. |
Grid search technique can be used for fair classification or fair regression. |
|
The meta algorithm here takes the fairness metric as part of the input and returns a classifier optimized w.r.t. |
|
Prejudice remover is an in-processing technique that adds a discrimination-aware regularization term to the learning objective. |
|
Fair K-Center Clustering inprocessing bias mitigation implements an approximation algorithm for the k-centers problem under the fairness contraint with running time linear in the size of the dataset and k (number of cluster). |
|
Fair K-median clustering inprocessing bias mitigation is an approximation algorithms for group representative k-median clustering. |
|
Fairlet Clustering inprocessing bias mitigation works in two steps: - The pointset is partitioned into subsets called fairlets that satisfy the fairness requirement and approximately preserve the k-median objective. |
|
Variational Fair Clustering helps you to find clusters with specified proportions of different demographic groups pertaining to a sensitive attribute of the dataset (group_a and group_b) for any well-known clustering method such as K-means, K-median or Spectral clustering (Normalized cut). |
Post-processing
Calibrated equalized odds postprocessing optimizes over calibrated classifier score outputs to find probabilities with which to change output labels with an equalized odds objective. |
|
Equalized odds postprocessing use linear programming to find the probability with which change favorable labels (y=1) to unfavorable labels (y=0) in the output estimator to optimize equalized odds. |
|
Reject option classification gives favorable outcomes (y=1) to unpriviliged groups and unfavorable outcomes (y=0)to priviliged groups in a confidence band around the decision boundary with the highest uncertainty. |
|
Linear Programmin Debiaser is a postprocessing algorithms designed to debias pretrained classifiers. |
|
Linear Programmin Debiaser is a postprocessing algorithms designed to debias pretrained classifiers. |
|
|
MLDebiaser postprocessing debias predictions w.r.t. |
|
Plugin Estimation and Calibration postprocessing optimizes over calibrated regressor outputs via a smooth optimization. |
Fair Regression with Wasserstein Barycenters learning a real-valued function that satisfies the Demographic Parity constraint. |