causal-fairness

This paper demonstrates that causal fairness mechanisms, specifically edge cuts on causal graphs, are portable across various synthetic data generator families including GANs and diffusion models, with minimal impact on data fidelity and utility.