Mitigation¶
Fairlearn contains the following algorithms for mitigating unfairness:
algorithm 
description 
binary classification 
regression 
supported fairness definitions 

A wrapper (reduction) approach to fair classification described in A Reductions Approach to Fair Classification 5. 
✔ 
✔ 
DP, EO, TPRP, FPRP, ERP, BGL 

A wrapper (reduction) approach described in Section 3.4 of A Reductions Approach to Fair Classification 5. For regression it acts as a gridsearch variant of the algorithm described in Section 5 of Fair Regression: Quantitative Definitions and Reductionbased Algorithms 4. 
✔ 
✔ 
DP, EO, TPRP, FPRP, ERP, BGL 

Postprocessing algorithm based on the paper Equality of Opportunity in Supervised Learning 6. This technique takes as input an existing classifier and the sensitive feature, and derives a monotone transformation of the classifier’s prediction to enforce the specified parity constraints. 
✔ 
✘ 
DP, EO, TPRP, FPRP 

Preprocessing algorithm that removes correlation between sensitive features and nonsensitive features through linear transformations. 
✔ 
✔ 
✘ 
DP refers to demographic parity, EO to equalized odds, TPRP to true positive rate parity, FPRP to false positive rate parity, ERP to error rate parity, and BGL to bounded group loss. For more information on the definitions refer to Fairness in Machine Learning. To request additional algorithms or fairness definitions, please open a new issue on GitHub.
Note
Fairlearn mitigation algorithms largely follow the
conventions of scikitlearn,
meaning that they implement the fit
method to train a model and the predict
method
to make predictions. However, in contrast with
scikitlearn,
Fairlearn algorithms can produce randomized predictors. Randomization of
predictions is required to satisfy many definitions of fairness. Because of
randomization, it is possible to get different outputs from the predictor’s
predict
method on identical data. For each of our algorithms, we provide
explicit access to the probability distribution used for randomization.
Preprocessing¶
Postprocessing¶
Reductions¶
On a high level, the reduction algorithms within Fairlearn enable unfairness mitigation for an arbitrary machine learning model with respect to userprovided fairness constraints. All of the constraints currently supported by reduction algorithms are groupfairness constraints. For more information on the supported fairness constraints refer to Fairness constraints for binary classification and Fairness constraints for regression.
Note
The choice of a fairness metric and fairness constraints is a crucial step in the AI development and deployment, and choosing an unsuitable constraint can lead to more harms. For a broader discussion of fairness as a sociotechnical challenge and how to view Fairlearn in this context refer to Fairness in Machine Learning.
The reductions approach for classification seeks to reduce binary
classification subject to fairness constraints to a sequence of weighted
classification problems (see 5), and similarly for regression (see 4).
As a result, the reduction algorithms
in Fairlearn only require a wrapper access to any “base” learning algorithm.
By this we mean that the “base” algorithm only needs to implement fit
and
predict
methods, as any standard scikitlearn estimator, but it
does not need to have any knowledge of the desired fairness constraints or sensitive features.
From an API perspective this looks as follows in all situations
>>> reduction = Reduction(base_estimator, constraints, **kwargs)
>>> reduction.fit(X_train, y_train, sensitive_features=sensitive_features)
>>> reduction.predict(X_test)
Fairlearn doesn’t impose restrictions on the referenced base_estimator
other than the existence of fit
and predict
methods.
At the moment, the base_estimator
’s fit
method also needs to
provide a sample_weight
argument which the reductions techniques use
to reweight samples.
In the future Fairlearn will provide functionality to handle this even
without a sample_weight
argument.
Before looking more into reduction algorithms, this section
reviews the supported fairness constraints. All of them
are expressed as objects inheriting from the base class Moment
.
Moment
’s main purpose is to calculate the constraint violation of a
current set of predictions through its gamma
function as well as to
provide signed_weights
that are used to relabel and reweight samples.
Fairness constraints for binary classification¶
All supported fairness constraints for binary classification inherit from
UtilityParity
. They are based on some underlying metric called
utility, which can be evaluated on individual data points and is averaged
over various groups of data points to form the utility parity constraint
of the form
where \(a\) is a sensitive feature value and \(e\) is an event identifier. Each data point has only one value of a sensitive feature, and belongs to at most one event. In many examples, there is only a single event \(*\), which includes all the data points. Other examples of events include \(Y=0\) and \(Y=1\). The utility parity requires that the mean utility within each event equals the mean utility of each group whose sensitive feature is \(a\) within that event.
The class UtilityParity
implements constraints that allow
some amount of violation of the utility parity constraints, where
the maximum allowed violation is specified either as a difference
or a ratio.
The differencebased relaxation starts out by representing the utility parity constraints as pairs of inequalities
and then replaces zero on the righthand side
with a value specified as difference_bound
. The resulting
constraints are instantiated as
>>> UtilityParity(difference_bound=0.01)
Note that satisfying these constraints does not mean
that the difference between the groups with the highest and
smallest utility in each event is bounded by difference_bound
.
The value of difference_bound
instead bounds
the difference between the utility of each group and the overall mean
utility within each event. This, however,
implies that the difference between groups in each event is
at most twice the value of difference_bound
.
The ratiobased relaxation relaxes the parity constraint as
for some value of \(r\) in (0,1]. For example, if \(r=0.9\), this means that within each event \(0.9 \cdot \text{utility}_{a,e} \leq \text{utility}_e\), i.e., the utility for each group needs to be at least 90% of the overall utility for the event, and \(0.9 \cdot \text{utility}_e \leq \text{utility}_{a,e}\), i.e., the overall utility for the event needs to be at least 90% of each group’s utility.
The two ratio constraints can be rewritten as
When instantiating the ratio constraints, we use ratio_bound
for \(r\),
and also allow further relaxation by replacing the zeros on the right hand side
by some nonnegative ratio_bound_slack
. The resulting instantiation
looks as
>>> UtilityParity(ratio_bound=0.9, ratio_bound_slack=0.01)
Similarly to the difference constraints, the ratio constraints do not directly bound the ratio between the pairs of groups, but such a bound is implied.
Note
It is not possible to specify both difference_bound
and
ratio_bound
for the same constraint object.
Demographic Parity¶
A binary classifier \(h(X)\) satisfies demographic parity if
In other words, the selection rate or percentage of samples with label 1 should be equal across all groups. Implicitly this means the percentage with label 0 is equal as well. In this case, the utility function is equal to \(h(X)\) and there is only a single event \(*\).
In the example below group "a"
has a selection rate of 60%,
"b"
has a selection rate of 20%. The overall selection rate is 40%,
so "a"
is 0.2 above the overall selection rate, and "b"
is
0.2 below. Invoking the method gamma
shows the values
of the lefthand sides of the constraints described
in Fairness constraints for binary classification, which is independent
of the provided difference_bound
. Note that the lefthand sides
corresponding to different values of sign
are just negatives
of each other.
The value of y_true
is in this example irrelevant to the calculations,
because the underlying utility in demographic parity, selection rate, does not
consider performance relative to the true labels, but rather proportions in
the predicted labels.
Note
When providing DemographicParity
to mitigation algorithms, only use
the constructor and the mitigation algorithm itself then invokes load_data
.
The example below uses load_data
to illustrate how DemographicParity
instantiates inequalities from Fairness constraints for binary classification.
>>> from fairlearn.reductions import DemographicParity
>>> from fairlearn.metrics import MetricFrame, selection_rate
>>> import numpy as np
>>> import pandas as pd
>>> dp = DemographicParity(difference_bound=0.01)
>>> X = np.array([[0], [1], [2], [3], [4], [5], [6], [7], [8], [9]])
>>> y_true = np.array([ 1 , 1 , 1 , 1 , 0, 0 , 0 , 0 , 0 , 0 ])
>>> y_pred = np.array([ 1 , 1 , 1 , 1 , 0, 0 , 0 , 0 , 0 , 0 ])
>>> sensitive_features = np.array(["a", "b", "a", "a", "b", "a", "b", "b", "a", "b"])
>>> selection_rate_summary = MetricFrame(metrics=selection_rate,
... y_true=y_true,
... y_pred=y_pred,
... sensitive_features=pd.Series(sensitive_features, name="SF 0"))
>>> selection_rate_summary.overall
0.4
>>> selection_rate_summary.by_group
SF 0
a 0.6
b 0.2
Name: selection_rate, dtype: object
>>> dp.load_data(X, y_true, sensitive_features=sensitive_features)
>>> dp.gamma(lambda X: y_pred)
sign event group_id
+ all a 0.2
b 0.2
 all a 0.2
b 0.2
dtype: float64
The ratio constraints for the demographic parity with ratio_bound
\(r\) (and ratio_bound_slack=0
) take form
Revisiting the same example as above we get
>>> dp = DemographicParity(ratio_bound=0.9, ratio_bound_slack=0.01)
>>> dp.load_data(X, y_pred, sensitive_features=sensitive_features)
>>> dp.gamma(lambda X: y_pred)
sign event group_id
+ all a 0.14
b 0.22
 all a 0.24
b 0.16
dtype: float64
Following the expressions for the lefthand sides of the constraints, we obtain
True Positive Rate Parity and False Positive Rate Parity¶
A binary classifier \(h(X)\) satisfies true positive rate parity if
and false positive rate parity if
In first case, we only have one event \(Y=1\) and ignore the samples with \(Y=0\), and in the second case vice versa. Refer to Equalized Odds for the fairness constraint type that simultaneously enforce both true positive rate parity and false positive rate parity by considering both events \(Y=0\) and \(Y=1\).
In practice this can be used in a differencebased relaxation as follows:
>>> from fairlearn.reductions import TruePositiveRateParity
>>> from fairlearn.metrics import true_positive_rate
>>> import numpy as np
>>> tprp = TruePositiveRateParity(difference_bound=0.01)
>>> X = np.array([[0], [1], [2], [3], [4], [5], [6], [7], [8], [9]])
>>> y_true = np.array([ 1 , 1 , 1 , 1 , 1, 1 , 1 , 0 , 0 , 0 ])
>>> y_pred = np.array([ 1 , 1 , 1 , 1 , 0, 0 , 0 , 1 , 0 , 0 ])
>>> sensitive_features = np.array(["a", "b", "a", "a", "b", "a", "b", "b", "a", "b"])
>>> tpr_summary = MetricFrame(metrics=true_positive_rate,
... y_true=y_true,
... y_pred=y_pred,
... sensitive_features=sensitive_features)
>>> tpr_summary.overall
0.5714285714285714
>>> tpr_summary.by_group
sensitive_feature_0
a 0.75...
b 0.33...
Name: true_positive_rate, dtype: object
>>> tprp.load_data(X, y_true, sensitive_features=sensitive_features)
>>> tprp.gamma(lambda X: y_pred)
sign event group_id
+ label=1 a 0.1785...
b 0.2380...
 label=1 a 0.1785...
b 0.2380...
dtype: float64
Note
When providing TruePositiveRateParity
or FalsePositiveRateParity
to mitigation algorithms, only use
the constructor. The mitigation algorithm itself then invokes load_data
.
The example uses load_data
to illustrate how TruePositiveRateParity
instantiates inequalities from Fairness constraints for binary classification.
Alternatively, a ratiobased relaxation is also available:
>>> tprp = TruePositiveRateParity(ratio_bound=0.9, ratio_bound_slack=0.01)
>>> tprp.load_data(X, y_true, sensitive_features=sensitive_features)
>>> tprp.gamma(lambda X: y_pred)
sign event group_id
+ label=1 a 0.1035...
b 0.2714...
 label=1 a 0.2357...
b 0.1809...
dtype: float64
Equalized Odds¶
A binary classifier \(h(X)\) satisfies equalized odds if it satisfies both true positive rate parity and false positive rate parity, i.e.,
The constraints represent the union of constraints for true positive rate and false positive rate.
>>> from fairlearn.reductions import EqualizedOdds
>>> eo = EqualizedOdds(difference_bound=0.01)
>>> eo.load_data(X, y_true, sensitive_features=sensitive_features)
>>> eo.gamma(lambda X: y_pred)
sign event group_id
+ label=0 a 0.3333...
b 0.1666...
label=1 a 0.1785...
b 0.2380...
 label=0 a 0.3333...
b 0.1666...
label=1 a 0.1785...
b 0.2380...
dtype: float64
Error Rate Parity¶
The error rate parity requires that the error rates should be the same across all groups. For a classifier \(h(X)\) this means that
In this case, the utility is equal to 1 if \(h(X)\ne Y\) and equal to
0 if \(h(X)=Y\), and so large value of utility here actually correspond
to poor outcomes. The differencebased relaxation specifies that
the error rate of any given group should not deviate from
the overall error rate by more than the value of difference_bound
.
>>> from fairlearn.reductions import ErrorRateParity
>>> from sklearn.metrics import accuracy_score
>>> accuracy_summary = MetricFrame(metrics=accuracy_score,
... y_true=y_true,
... y_pred=y_pred,
... sensitive_features=sensitive_features)
>>> accuracy_summary.overall
0.6
>>> accuracy_summary.by_group
sensitive_feature_0
a 0.8
b 0.4
Name: accuracy_score, dtype: object
>>> erp = ErrorRateParity(difference_bound=0.01)
>>> erp.load_data(X, y_true, sensitive_features=sensitive_features)
>>> erp.gamma(lambda X: y_pred)
sign event group_id
+ all a 0.2
b 0.2
 all a 0.2
b 0.2
dtype: float64
Note
When providing ErrorRateParity
to mitigation algorithms, only use
the constructor. The mitigation algorithm itself then invokes load_data
.
The example uses load_data
to illustrate how ErrorRateParity
instantiates inequalities from Fairness constraints for binary classification.
Alternatively, error rate parity can be relaxed via ratio constraints as
with a ratio_bound
\(r\). The usage is identical with other
constraints:
>>> from fairlearn.reductions import ErrorRateParity
>>> erp = ErrorRateParity(ratio_bound=0.9, ratio_bound_slack=0.01)
>>> erp.load_data(X, y_true, sensitive_features=sensitive_features)
>>> erp.gamma(lambda X: y_pred)
sign event group_id
+ all a 0.22
b 0.14
 all a 0.16
b 0.24
dtype: float64
Control features¶
The above examples of Moment
(Demographic Parity,
True and False Positive Rate Parity,
Equalized Odds and Error Rate Parity) all support the concept
of control features when applying their fairness constraints.
A control feature stratifies the dataset, and applies the fairness constraint
within each stratum, but not between strata.
One case this might be useful is a loan scenario, where we might want
to apply a mitigation for the sensitive features while controlling for some
other feature(s).
This should be done with caution, since the control features may have a
correlation with the sensitive features due to historical biases.
In the loan scenario, we might choose to control for income level, on the
grounds that higher income individuals are more likely to be able to repay
a loan.
However, due to historical bias, there is a correlation between the income level
of individuals and their race and gender.
Control features modify the above equations. Consider a control feature value, drawn from a set of valid values (that is, \(c \in \mathcal{C}\)). The equation given above for Demographic Parity will become:
The other constraints acquire similar modifications.
Fairness constraints for multiclass classification¶
Reductions approaches do not support multiclass classification yet at this point. If this is an important scenario for you please let us know!
Fairness constraints for regression¶
The performance objective in the regression scenario is to minimize the
loss of our regressor \(h\). The loss can be expressed as
SquareLoss
or AbsoluteLoss
. Both take constructor arguments
min_val
and max_val
that define the value range within which
the loss is evaluated. Values outside of the value range get clipped.
>>> from fairlearn.reductions import SquareLoss, AbsoluteLoss, ZeroOneLoss
>>> y_true = [0, 0.3, 1, 0.9]
>>> y_pred = [0.1, 0.2, 0.9, 1.3]
>>> SquareLoss(0, 2).eval(y_true, y_pred)
array([0.01, 0.01, 0.01, 0.16])
>>> # clipping at 1 reduces the error for the fourth entry
>>> SquareLoss(0, 1).eval(y_true, y_pred)
array([0.01, 0.01, 0.01, 0.01])
>>> AbsoluteLoss(0, 2).eval(y_true, y_pred)
array([0.1, 0.1, 0.1, 0.4])
>>> AbsoluteLoss(0, 1).eval(y_true, y_pred)
array([0.1, 0.1, 0.1, 0.1])
>>> # ZeroOneLoss is identical to AbsoluteLoss(0, 1)
>>> ZeroOneLoss().eval(y_true, y_pred)
array([0.1, 0.1, 0.1, 0.1])
When using Fairlearn’s reduction techniques for regression it’s required to
specify the type of loss by passing the corresponding loss object when
instantiating the object that represents our fairness constraint. The only
supported type of constraint at this point is BoundedGroupLoss
.
Bounded Group Loss¶
Bounded group loss requires the loss of each group to be below a userspecified amount \(\zeta\). If \(\zeta\) is chosen reasonably small the losses of all groups are very similar. Formally, a predictor \(h\) satisfies bounded group loss at level \(\zeta\) under a distribution over \((X, A, Y)\) if
In the example below we use BoundedGroupLoss
with
ZeroOneLoss
on two groups "a"
and "b"
.
Group "a"
has an average loss of \(0.05\), while group
"b"
’s average loss is \(0.5\).
>>> from fairlearn.reductions import BoundedGroupLoss, ZeroOneLoss
>>> from sklearn.metrics import mean_absolute_error
>>> bgl = BoundedGroupLoss(ZeroOneLoss(), upper_bound=0.1)
>>> X = np.array([[0], [1], [2], [3]])
>>> y_true = np.array([0.3, 0.5, 0.1, 1.0])
>>> y_pred = np.array([0.3, 0.6, 0.6, 0.5])
>>> sensitive_features = np.array(["a", "a", "b", "b"])
>>> mae_frame = MetricFrame(metrics=mean_absolute_error,
... y_true=y_true,
... y_pred=y_pred,
... sensitive_features=pd.Series(sensitive_features, name="SF 0"))
>>> mae_frame.overall
0.275
>>> mae_frame.by_group
SF 0
a 0.05
b 0.5
Name: mean_absolute_error, dtype: object
>>> bgl.load_data(X, y_true, sensitive_features=sensitive_features)
>>> bgl.gamma(lambda X: y_pred)
group_id
a 0.05
b 0.50
Name: loss, dtype: float64
Note
In the example above the BoundedGroupLoss
object does not use the
upper_bound
argument. It is only used by reductions techniques
during the unfairness mitigation. As a result the constraint violation
detected by gamma
is identical to the mean absolute error.
Exponentiated Gradient¶
Grid Search¶
References:
 4(1,2)
Agarwal, Dudik, Wu “Fair Regression: Quantitative Definitions and Reductionbased Algorithms”, ICML, 2019.
 5(1,2,3)
Agarwal, Beygelzimer, Dudik, Langford, Wallach “A Reductions Approach to Fair Classification”, ICML, 2018.
 6
Hardt, Price, Srebro “Equality of Opportunity in Supervised Learning”, NeurIPS, 2016.