Sequential Conditional Transport on Probabilistic Graphs for Interpretable Counterfactual Fairness
This work addresses fairness at the individual level for applications in machine learning, but it appears incremental as it combines and extends existing approaches.
The paper tackles the problem of generating interpretable counterfactuals for fairness by linking causal graph adaptations and optimal transport, extending Knothe's rearrangement and triangular transport to probabilistic graphical models. It demonstrates the method through numerical experiments on synthetic and real datasets, though no concrete numbers are provided in the abstract.
In this paper, we link two existing approaches to derive counterfactuals: adaptations based on a causal graph, and optimal transport. We extend "Knothe's rearrangement" and "triangular transport" to probabilistic graphical models, and use this counterfactual approach, referred to as sequential transport, to discuss fairness at the individual level. After establishing the theoretical foundations of the proposed method, we demonstrate its application through numerical experiments on both synthetic and real datasets.