mlsanityData to 14 Oct 2026

Expert-Aided Causal Discovery of Ancestral Graphs

arXiv 2309.12032 ↗First posted 21 Sep 2023cs.LG stat.MLcited by 8

Not on the map. Why?

Abstract (arXiv)

Causal discovery (CD) is an important component of many scientific applications, yet most techniques produce unreliable point estimates that often contradict expert knowledge. To mitigate this, recent research has focused on ex-ante incorporation of background knowledge into the CD process, typically under an unrealistic causal sufficiency assumption. When probing experts is costly (e.g., hidden behind expensive LLM APIs), however, ex-post model refinement that maximizes query utility is preferable. Also, when independent experts provide conflicting but better-than-random feedback, a principled aggregation method is required. In this context, we introduce the first CD algorithm that enables (i) distributional inference over ancestral graphs (AGs), which represent causal systems under latent confounding, and (ii) integration of both ex-ante and uncertain ex-post expert knowledge. Briefly, our method is a diversity-seeking reinforcement learning algorithm, termed Ancestral GFlowNet (AGFN), whose policy we iteratively refine based on a Bayesian model of the noisy expert feedback. Importantly, we prove convergence to the true AG given sufficiently accurate responses. Through validation on synthetic and realistic datasets using simulated humans and LLMs, we show AGFN is competitive with or superior to strong baselines in terms of structural Hamming distance and Bayesian Information Criterion.

Key facts. Cited by 8 papers in the map (data through 14 Oct 2026). Builds on 33 of its 107 references that are in the map. First posted on arXiv 21 Sep 2023.

Builds on

33 of 107 references.

References · 33

  1. 2014Adam: A Method for Stochastic Optimizationcited by 37,240
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  3. 2018How Powerful are Graph Neural Networks?cited by 4,393
  4. 2023Can LLMs Express Their Uncertainty? An Empirical Evaluation of Confidence Elicitation in LLMscited by 974
  5. 2021Flow Network based Generative Models for Non-Iterative Diverse Candidate Generationcited by 392
  6. 2023Causal Reasoning and Large Language Models: Opening a New Frontier for Causalitycited by 353
  7. 2022Demonstrate-Search-Predict: Composing retrieval and language models for knowledge-intensive NLPcited by 318
  8. 2021GFlowNet Foundationscited by 290
  9. 2022Trajectory balance: Improved credit assignment in GFlowNetscited by 251
  10. 2022Biological Sequence Design with GFlowNetscited by 174
  11. 2022Bayesian Structure Learning with Generative Flow Networkscited by 153
  12. 2022Generative Flow Networks for Discrete Probabilistic Modelingcited by 114
21 more
  1. 2023A theory of continuous generative flow networkscited by 101
  2. 2023Better Training of GFlowNets with Local Credit and Incomplete Trajectoriescited by 85
  3. 2023Towards Understanding and Improving GFlowNet Trainingcited by 75
  4. 2023Robust Scheduling with GFlowNetscited by 69
  5. 2020Differentiable Causal Discovery Under Unmeasured Confoundingcited by 56
  6. 2022Generative Augmented Flow Networkscited by 50
  7. 2023Joint Bayesian Inference of Graphical Structure and Parameters with a Single Generative Flow Networkcited by 50
  8. 2013Learning Sparse Causal Models is not NP-hardcited by 49
  9. 2013Strong Completeness and Faithfulness in Bayesian Networkscited by 45
  10. 2023GFlowNet-EM for learning compositional latent variable modelscited by 40
  11. 2022Causal Effect Identification in Cluster DAGscited by 36
  12. 2023DynGFN: Towards Bayesian Inference of Gene Regulatory Networks with GFlowNetscited by 27
  13. 2019Ordering-Based Causal Structure Learning in the Presence of Latent Variablescited by 26
  14. 2016Ancestral Causal Inferencecited by 24
  15. 2022GFlowOut: Dropout with Generative Flow Networkscited by 22
  16. 2022Reliable Causal Discovery with Improved Exact Search and Weaker Assumptionscited by 19
  17. 2023Causal Reasoning in the Presence of Latent Confounders via Neural ADMG Learningcited by 15
  18. 2015Learning Structures of Bayesian Networks for Variable Groupscited by 12
  19. 2021Typing assumptions improve identification in causal discoverycited by 10
  20. 2015Distributional Equivalence and Structure Learning for Bow-free Acyclic Path Diagramscited by 8
  21. 2023GFlowNets with Human Feedbackcited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 3; 2025: 2; 2026: 3.

8 papers.

  1. 2024Large Language Models for Constrained-Based Causal Discoverycited by 15
  2. 2024On Divergence Measures for Training GFlowNetscited by 6
  3. 2024Embarrassingly Parallel GFlowNetscited by 3
  4. 2025Learning to Defer for Causal Discovery with Imperfect Expertscited by 2
  5. 2025Sequential Causal Discovery with Noisy Language Model Priorscited by 0
  6. 2026Causal Preference Elicitationcited by 0
  7. 2026CausalSteward: An Agentic Divide-Conquer-Combine Copilot for Causal Discoverycited by 0
  8. 2026Particle GFlowNets: Rethinking Generative Marginalization Modelscited by 0