Modeling implicit bias with fuzzy cognitive maps

This paper presents a Fuzzy Cognitive Map model to quantify implicit bias in structured datasets where features can be numeric or discrete. In our proposal, problem are mapped neural concepts that initially activated by experts when running what-if simulations, whereas weights connecting the represent absolute correlation/association patterns between features. addition, we introduce new reasoning mechanism equipped with normalization-like transfer function prevents neurons from saturating. Another advantage of this is it easily controlled regulating nonlinearity updating neurons’ activation values each iteration. Finally, study convergence and derive analytical conditions concerning existence unicity fixed-point attractors.