mlsanityData to 14 Oct 2026

Unsupervised Causal Abstractions Discovery

arXiv 2606.19594 ↗First posted 17 Jun 2026cs.LGcited by 1

Not on the map. Why?

Abstract (arXiv)

Causal abstractions formalize when a high-level structural causal model (SCM) captures the interventional behavior of a lower-level SCM. Existing applications of this notion largely follow a hypothesis-testing paradigm: an expert proposes a candidate high-level model and then evaluates if the low-level system implements it. We study the complementary problem of learning a high-level model directly from low-level measurements. Our contributions leverage hypotheses from low-rank causal discovery, and can be summarized as follows: (1) we show that observations generated by a low-rank graph induce latents that form a causal abstraction, (2) we provide identifiability results about these latents, and (3) we propose a practical objective to learn this high-level SCM.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Builds on 26 of its 45 references that are in the map. First posted on arXiv 17 Jun 2026.

Builds on

26 of 45 references.

References · 26

  1. 2016The Concrete Distribution: A Continuous Relaxation of Discrete Random Variablescited by 1,148
  2. 2018DAGs with NO TEARS: Continuous Optimization for Structure Learningcited by 752
  3. 2021Causal Abstractions of Neural Networkscited by 422
  4. 2023Finding Alignments Between Interpretable Causal Variables and Distributed Neural Representationscited by 209
  5. 2023Causal Abstraction: A Theoretical Foundation for Mechanistic Interpretabilitycited by 205
  6. 2017Causal Consistency of Structural Equation Modelscited by 118
  7. 2018Abstracting Causal Modelscited by 107
  8. 2023Interpretability at Scale: Identifying Causal Mechanisms in Alpacacited by 103
  9. 2021Identifiable Deep Generative Models via Sparse Decodingcited by 54
  10. 2023Joint Bayesian Inference of Graphical Structure and Parameters with a Single Generative Flow Networkcited by 50
  11. 2022Large-Scale Differentiable Causal Discovery of Factor Graphscited by 47
  12. 2019Scaling structural learning with NO-BEARS to infer causal transcriptome networkscited by 39
14 more
  1. 2023Stable Differentiable Causal Discoverycited by 31
  2. 2015Multi-Level Cause-Effect Systemscited by 30
  3. 2013The Randomized Dependence Coefficientcited by 26
  4. 2020On Low Rank Directed Acyclic Graphs and Causal Structure Learningcited by 23
  5. 2023Jointly Learning Consistent Causal Abstractions Over Multiple Interventional Distributionscited by 23
  6. 2023Targeted Reduction of Causal Modelscited by 15
  7. 2023Causal Optimal Transport of Abstractionscited by 13
  8. 2024Learning Causal Abstractions of Linear Structural Causal Modelscited by 11
  9. 2025Cooper: A Library for Constrained Optimization in Deep Learningcited by 11
  10. 2025Causal Abstraction Learning based on the Semantic Embedding Principlecited by 9
  11. 2022From graphs to DAGs: a low-complexity model and a scalable algorithmcited by 5
  12. 2023Neuro-Causal Factor Analysiscited by 4
  13. 2026Structural Causal Bottleneck Modelscited by 4
  14. 2025Deep Discrete Encoders: Identifiable Deep Generative Models for Rich Data with Discrete Latent Layerscited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026When Do Learned Blocks Preserve Causal Meaning? Block Identifiability and Constructive Causal Abstractionscited by 0