Mining bias-target Alignment from Voronoi Cells
This addresses fairness and generalization issues in AI for applications where bias mitigation is critical, though it appears incremental as it builds on existing debiasing research.
The paper tackles the problem of bias in deep neural networks by proposing a bias-agnostic method that quantifies bias-target alignment to discourage bias propagation, achieving comparable performance to state-of-the-art supervised approaches on common debiasing datasets.
Despite significant research efforts, deep neural networks are still vulnerable to biases: this raises concerns about their fairness and limits their generalization. In this paper, we propose a bias-agnostic approach to mitigate the impact of bias in deep neural networks. Unlike traditional debiasing approaches, we rely on a metric to quantify ``bias alignment/misalignment'' on target classes, and use this information to discourage the propagation of bias-target alignment information through the network. We conduct experiments on several commonly used datasets for debiasing and compare our method to supervised and bias-specific approaches. Our results indicate that the proposed method achieves comparable performance to state-of-the-art supervised approaches, although it is bias-agnostic, even in presence of multiple biases in the same sample.