Assessing Classifier Fairness with Collider Bias
This addresses fairness issues in algorithmic decision-making for audit agencies, but it is incremental as it builds on existing fairness assessment methods.
The paper tackles the problem of collider bias in fairness assessment of machine learning classifiers, developing theorems and an algorithm that significantly reduces such biases in experiments and simulations.
The increasing application of machine learning techniques in everyday decision-making processes has brought concerns about the fairness of algorithmic decision-making. This paper concerns the problem of collider bias which produces spurious associations in fairness assessment and develops theorems to guide fairness assessment avoiding the collider bias. We consider a real-world application of auditing a trained classifier by an audit agency. We propose an unbiased assessment algorithm by utilising the developed theorems to reduce collider biases in the assessment. Experiments and simulations show the proposed algorithm reduces collider biases significantly in the assessment and is promising in auditing trained classifiers.