mlsanityData to 14 Oct 2026

Differentiable Cyclic Causal Discovery Under Unmeasured Confounders

arXiv 2508.08450 ↗First posted 11 Aug 2025cs.LG stat.ME stat.MLcited by 4

Not on the map. Why?

Abstract (arXiv)

Understanding causal relationships between variables is fundamental across scientific disciplines. Most causal discovery algorithms rely on two key assumptions: (i) all variables are observed, and (ii) the underlying causal graph is acyclic. While these assumptions simplify theoretical analysis, they are often violated in real-world systems, such as biological networks. Existing methods that account for confounders either assume linearity or struggle with scalability. To address these limitations, we propose DCCD-CONF, a novel framework for differentiable learning of nonlinear cyclic causal graphs in the presence of unmeasured confounders using interventional data. Our approach alternates between optimizing the graph structure and estimating the confounder distribution by maximizing the log-likelihood of the data. Through experiments on synthetic data and real-world gene perturbation datasets, we show that DCCD-CONF outperforms state-of-the-art methods in both causal graph recovery and confounder identification. Additionally, we also provide consistency guarantees for our framework, reinforcing its theoretical soundness.

Key facts. Cited by 4 papers in the map (data through 14 Oct 2026). Builds on 31 of its 65 references that are in the map. First posted on arXiv 11 Aug 2025.

Builds on

31 of 65 references.

References · 31

  1. 2016Categorical Reparameterization with Gumbel-Softmaxcited by 2,739
  2. 2018DAGs with NO TEARS: Continuous Optimization for Structure Learningcited by 752
  3. 2018Invertible Residual Networkscited by 365
  4. 2019DAG-GNN: DAG Structure Learning with Graph Neural Networkscited by 331
  5. 2019Residual Flows for Invertible Generative Modelingcited by 257
  6. 2020Differentiable Causal Discovery from Interventional Datacited by 218
  7. 2019Learning Sparse Nonparametric DAGscited by 207
  8. 2020On the Role of Sparsity and DAG Constraints for Learning Linear DAGscited by 191
  9. 2017Invariant Causal Prediction for Nonlinear Modelscited by 140
  10. 2022DAGMA: Learning DAGs via M-matrices and a Log-Determinant Acyclicity Characterizationcited by 140
  11. 2016Foundations of Structural Causal Models with Cycles and Latent Variablescited by 134
  12. 2019Masked Gradient-Based Causal Structure Learningcited by 89
19 more
  1. 2017Permutation-based Causal Inference Algorithms with Interventionscited by 78
  2. 2020Differentiable Causal Discovery Under Unmeasured Confoundingcited by 56
  3. 2014Constraint-based Causal Discovery from Multiple Interventions over Overlapping Variable Setscited by 48
  4. 2022Large-Scale Differentiable Causal Discovery of Factor Graphscited by 47
  5. 2016Double/Debiased Machine Learning for Treatment and Causal Parameterscited by 43
  6. 2019Scaling structural learning with NO-BEARS to infer causal transcriptome networkscited by 39
  7. 2018Constraint-based Causal Discovery for Non-Linear Structural Causal Models with Cycles and Latent Confounderscited by 35
  8. 2013A Discovery Algorithm for Directed Cyclic Graphscited by 34
  9. 2017Markov Properties for Graphical Models with Cycles and Latent Variablescited by 32
  10. 2013Directed Cyclic Graphical Representations of Feedback Modelscited by 28
  11. 2021Implicit Normalizing Flowscited by 28
  12. 2013Cyclic Causal Discovery from Continuous Equilibrium Datacited by 21
  13. 2023NODAGS-Flow: Nonlinear Cyclic Causal Structure Learningcited by 20
  14. 2021Integer Programming for Causal Structure Learning in the Presence of Latent Variablescited by 15
  15. 2020Structure Learning for Cyclic Linear Causal Modelscited by 13
  16. 2014A factorization criterion for acyclic directed mixed graphscited by 8
  17. 2024Doubly Robust Inference in Causal Latent Factor Modelscited by 4
  18. 2024Generalization of LiNGAM that allows confoundingcited by 2
  19. 2025ExMAG: Learning of Maximally Ancestral Graphscited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 4.

4 papers.

  1. 2026Consistency evaluation of benchmarks used for causal discoverycited by 1
  2. 2026RECLAIM: Cyclic Causal Discovery Amid Measurement Noisecited by 0
  3. 2026SCOUT: Cyclic Causal Discovery Under Soft Interventions with Unknown Targetscited by 0
  4. 2026Differentiable Structure Learning for Cyclic Linear Gaussian Models with Latent Confounderscited by 0