Representation-based fairness evaluation and bias correction robustness assessment in neural networks

While machine learning has achieved high predictive performance in many domains, decisions may still be biased and unfair regarding specific demographic groups characterized by sensitive attributes such as gender, age, or race. In this paper, we introduce a novel approach to assess model fairness and bias correction robustness based on Computational Profile Distance (CPD) analysis with respect to sensitive attributes. To study model fairness, we quantify the model’s representation difference using the computational profile learned from different subgroups (e.g., male and female) on the individual and group level. To analyze the robustness of bias correction outcomes, we compare the correction suggestions provided based on confidence (i.e., softmax score) and likelihood (i.e., CPD). To demonstrate the potential of the proposed approach, experiments have been performed using 24 models targeting 3 datasets used in previous fairness studies. Our experiments showed that computational profile distributions can effectively address model fairness from a representation perspective. Further, the experiments indicated that confidence-based bias correction decisions can vary largely from likelihood-based ones, and we should take both suggestions into account to obtain robust outcomes. Demonstrated with a set of experiments, our CPD-based approaches can help users build their trust in fairness assessment and bias mitigation of AI decisions, in ethically sensitive domains such as human resources, finance, health, and more. • Fairness estimation from a representation perspective using limited data. • Combine confidence and likelihood metrics to obtain robust bias corrections. • Visualizations providing support for practical usage.