mlsanityData to 7 Oct 2026

Local Causal Discovery for Structural Evidence of Direct Discrimination

arXiv 2405.14848 ↗First posted 23 May 2024stat.ML cs.LGcited by 6

Not on the map. Why?

Abstract (arXiv)

Identifying the causal pathways of unfairness is a critical objective for improving policy design and algorithmic decision-making. Prior work in causal fairness analysis often requires knowledge of the causal graph, hindering practical applications in complex or low-knowledge domains. Moreover, global discovery methods that learn causal structure from data can display unstable performance on finite samples, preventing robust fairness conclusions. To mitigate these challenges, we introduce local discovery for direct discrimination (LD3): a method that uncovers structural evidence of direct unfairness by identifying the causal parents of an outcome variable. LD3 performs a linear number of conditional independence tests relative to variable set size, and allows for latent confounding under the sufficient condition that all parents of the outcome are observed. We show that LD3 returns a valid adjustment set (VAS) under a new graphical criterion for the weighted controlled direct effect, a qualitative indicator of direct discrimination. LD3 limits unnecessary adjustment, providing interpretable VAS for assessing unfairness. We use LD3 to analyze causal fairness in two complex decision systems: criminal recidivism prediction and liver transplant allocation. LD3 was more time-efficient and returned more plausible results on real-world data than baselines, which took 46× to 5870× longer to execute.

Key facts. Cited by 6 papers in the map (data through 7 Oct 2026). Builds on 29 of its 114 references that are in the map. First posted on arXiv 23 May 2024.

Builds on

29 of 114 references.

References · 29

  1. 2018DAGs with NO TEARS: Continuous Optimization for Structure Learningcited by 745
  2. 2013Direct and Indirect Effectscited by 276
  3. 2013Causal Discovery with Continuous Additive Noise Modelscited by 252
  4. 2017Avoiding Discrimination through Causal Reasoningcited by 191
  5. 2018The Measure and Mismeasure of Fairnesscited by 94
  6. 2016A causal framework for discovering and removing direct and indirect discriminationcited by 71
  7. 2017Conditional independence testing based on a nearest-neighbor estimator of conditional mutual informationcited by 70
  8. 2020Survey on Causal-based Machine Learning Fairness Notionscited by 53
  9. 2013Learning Sparse Causal Models is not NP-hardcited by 48
  10. 2023Causal Discovery with Score Matching on Additive Models with Arbitrary Noisecited by 41
  11. 2021Fairness Perceptions of Algorithmic Decision-Making: A Systematic Review of the Empirical Literaturecited by 27
  12. 2018A Unified View of Causal and Non-causal Feature Selectioncited by 23
17 more
  1. 2013On the definition of a confoundercited by 19
  2. 2021Necessary and sufficient graphical conditions for optimal adjustment sets in causal graphical models with hidden variablescited by 16
  3. 2022Causal Discovery for Fairnesscited by 16
  4. 2023On Learning Necessary and Sufficient Causal Graphscited by 14
  5. 2018cuPC: CUDA-based Parallel PC Algorithm for Causal Structure Learning on GPUcited by 12
  6. 2020Causal Feature Selection via Orthogonal Searchcited by 11
  7. 2021Finding Valid Adjustments under Non-ignorability with Minimal DAG Knowledgecited by 10
  8. 2023Local Causal Discovery for Estimating Causal Effectscited by 10
  9. 2020Local search for efficient causal effect estimationcited by 9
  10. 2023Local Discovery by Partitioning: Polynomial-Time Causal Discovery Around Exposure-Outcome Pairscited by 9
  11. 2023Causal Fairness for Outcome Controlcited by 8
  12. 2020Insights into the "cross-world" independence assumption of causal mediation analysiscited by 7
  13. 2024Local Causal Discovery with Linear non-Gaussian Cyclic Modelscited by 7
  14. 2015Exposing the Probabilistic Causal Structure of Discriminationcited by 5
  15. 2022Confounder Selection: Objectives and Approachescited by 4
  16. 2023Front-door Adjustment Beyond Markov Equivalence with Limited Graph Knowledgecited by 4
  17. 2023A Causal Framework for Decomposing Spurious Variationscited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 1; 2025: 0; 2026: 5.

6 papers.

  1. 2024Local Learning for Covariate Selection in Nonparametric Causal Effect Estimation with Latent Variablescited by 3
  2. 2026Data Augmentation via Causal-Residual Bootstrappingcited by 0
  3. 2026From Local to Cluster: A Unified Framework for Causal Discovery with Latent Variablescited by 0
  4. 2026Data-Driven Covariate Selection for Nonparametric and Cycle-Agnostic Causal Effect Estimationcited by 0
  5. 2026Local Covariate Selection for Average Causal Effect Estimation without Pretreatment and Causal Sufficiency Assumptionscited by 0
  6. 2026Integrating Background Knowledge for Scalable Causal Discoverycited by 0