Sequential Conditional Transport on Probabilistic Graphs for Interpretable Counterfactual Fairness

Agathe Fernandes Machado, Arthur Charpentier, Ewen Gallic
8/6/2024
cs.LGstat.ME

No official code found. Best implementation: fer-agathe/sequential_transport (0 stars, MIT, last commit 2024-12-25).

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.

View on arXivView PDF

Code Implementations(1)

fer-agathe/sequential_transportCommunity100%
MIT
00R, LuaJun 28, 20241 years agoMIT

Similar Papers

Optimal Transport on Categorical Data for Counterfactuals using Compositional Data and Dirichlet Transport

87% match1 repo2025

What is a good matching of probability measures? A counterfactual lens on transport maps

83% match2025

Achieving Counterfactual Fairness with Imperfect Structural Causal Model

80% match2023

Counterfactually Fair Reinforcement Learning via Sequential Data Preprocessing

78% match2025

Mitigating topology biases in Graph Diffusion via Counterfactual Intervention

78% match2026

Learning Causal Graphs via Monotone Triangular Transport Maps

78% match2023

Cite this paper

@article{machado2024sequential,
  title  = {Sequential Conditional Transport on Probabilistic Graphs for Interpretable Counterfactual Fairness},
  author = {Agathe Fernandes Machado and Arthur Charpentier and Ewen Gallic},
  year   = {2024},
  eprint = {2408.03425},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2408.03425v3}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2408.03425.svg)](https://codeofpaper.com/paper/2408.03425)

Discussion