mlsanityData to 7 Oct 2026

xWhyL: Causal Interactive Learning

arXiv 2609.26037 ↗First posted 22 Sep 2026cs.LG cs.AI stat.MLcited by 0

On the map: LLM reasoning → Causal reasoning in LLMs

Abstract (arXiv)

Explanations are central to causal reasoning, and cognitive science has long established that the human drive to explain is itself a mechanism for learning about causality. Despite this, learning from those abductive signals is largely ignored in artificial intelligence. While explainable AI (XAI) increasingly draws on causal models to generate explanations, the converse direction about what explanations can do for causality remains largely unexplored. To fill this gap, we propose xWhyL, a formal framework connecting causality and XAI by learning causal models from explanations. We develop a mathematical theory that translates explanations into a learning signal complementary to observational data, and demonstrate how it enables overcoming the limits of observational causal discovery. As explanations can be derived from incorrect beliefs and clash with data, a tension we call the Causal Tug-of-War, we prove conditions under which our framework rejects misspecified explanations rather than absorbing them. Our practical instantiation, Causal Interactive Learning (CIL), shows how expert explanations can efficiently support causal discovery and distinguish correct from incorrect explanations.

Key facts. No citing papers in the map yet (data through 7 Oct 2026). Placed in Causal reasoning in LLMs, one of 17 directions in LLM reasoning. Builds on 30 of its 59 references that are in the map. First posted on arXiv 22 Sep 2026.

Builds on

30 of 59 references.

References · 30

  1. 2018A Survey Of Methods For Explaining Black Box Modelscited by 970
  2. 2018DAGs with NO TEARS: Continuous Optimization for Structure Learningcited by 745
  3. 2017Right for the Right Reasons: Training Differentiable Models by Constraining their Explanationscited by 293
  4. 2020Differentiable Causal Discovery from Interventional Datacited by 218
  5. 2020On the Role of Sparsity and DAG Constraints for Learning Linear DAGscited by 186
  6. 2013Causal Inference and Causal Explanation with Background Knowledgecited by 168
  7. 2021Beware of the Simulated DAG! Causal Discovery Benchmarks May Be Easy To Gamecited by 159
  8. 2022Bayesian Structure Learning with Generative Flow Networkscited by 153
  9. 2020Making deep neural networks right for the right scientific reasons by interacting with their explanationscited by 141
  10. 2022DAGMA: Learning DAGs via M-matrices and a Log-Determinant Acyclicity Characterizationcited by 135
  11. 2021DiBS: Differentiable Bayesian Structure Learningcited by 105
  12. 2020Right for the Right Concept: Revising Neuro-Symbolic Concepts by Interacting with their Explanationscited by 94
18 more
  1. 2018Characterizing and Learning Equivalence Classes of Causal DAGs under Interventionscited by 76
  2. 2021BCD Nets: Scalable Variational Approaches for Bayesian Causal Discoverycited by 68
  3. 2022TalkToModel: Explaining Machine Learning Models with Interactive Natural Language Conversationscited by 65
  4. 2022Leveraging Explanations in Interactive Machine Learning: An Overviewcited by 52
  5. 2015High-dimensional consistency in score-based and hybrid structure learningcited by 44
  6. 2021Tell me why! Explanations support learning relational and causal structurecited by 36
  7. 2021The impact of prior knowledge on causal structure learningcited by 35
  8. 2023BayesDAG: Gradient-Based Posterior Inference for Causal Discoverycited by 33
  9. 2019Neuropathic Pain Diagnosis Simulator for Causal Discovery Algorithm Evaluationcited by 32
  10. 2013Theory Refinement on Bayesian Networkscited by 28
  11. 2023Learning by Self-Explainingcited by 17
  12. 2025Extremely Greedy Equivalence Searchcited by 8
  13. 2024Markov Equivalence and Consistency in Differentiable Structure Learningcited by 6
  14. 2024Local Causal Discovery with Background Knowledgecited by 5
  15. 2019ASP-based Discovery of Semi-Markovian Causal Models under Weaker Assumptionscited by 3
  16. 2025Learning to Defer for Causal Discovery with Imperfect Expertscited by 2
  17. 2025Robust Causal Discovery under Imperfect Structural Constraintscited by 2
  18. 2025Less Greedy Equivalence Searchcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Explainable AI: Learning from the Learners
  2. 2025Towards Error-Centric Intelligence II: Energy-Structured Causal Models
  3. 2023Explainable Reinforcement Learning via a Causal World Model
  4. 2025Causal-Copilot: An Autonomous Causal Analysis Agent
  5. 2025LLM Cannot Discover Causality, and Should Be Restricted to Non-Decisional Support in Causal Discovery
  6. 2026CauSim: Scaling Causal Reasoning with Increasingly Complex Causal Simulators
  7. 2026CausalSteward: An Agentic Divide-Conquer-Combine Copilot for Causal Discovery
  8. 2025Causal Reasoning in Pieces: Modular In-Context Learning for Causal Discovery