mlsanityData to 14 Oct 2026

Counterfactuals uncover the modular structure of deep generative models

arXiv 1812.03253 ↗First posted 8 Dec 2018cs.LG stat.MLcited by 51

Not on the map. Why?

Abstract (arXiv)

Deep generative models can emulate the perceptual properties of complex image datasets, providing a latent representation of the data. However, manipulating such representation to perform meaningful and controllable transformations in the data space remains challenging without some form of supervision. While previous work has focused on exploiting statistical independence to disentangle latent factors, we argue that such requirement is too restrictive and propose instead a non-statistical framework that relies on counterfactual manipulations to uncover a modular structure of the network composed of disentangled groups of internal variables. Experiments with a variety of generative models trained on complex image datasets show the obtained modules can be used to design targeted interventions. This opens the way to applications such as computationally efficient style transfer and the automated assessment of robustness to contextual changes in pattern recognition systems.

Key facts. Cited by 51 papers in the map (data through 14 Oct 2026). First posted on arXiv 8 Dec 2018.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 19; 2022: 14; 2023: 10; 2024: 6; 2025: 1; 2026: 1.

51 papers.

Sort:

By area:

  1. 2021Towards Causal Representation Learningcited by 209
  2. 2023Causal Abstraction: A Theoretical Foundation for Mechanistic Interpretabilitycited by 205
  3. 2021Counterfactual Generative Networkscited by 111
  4. 2021Inducing Causal Structure for Interpretable Neural Networkscited by 110
  5. 2021Independent mechanism analysis, a new concept?cited by 97
  6. 2021Counterfactual Zero-Shot and Open-Set Visual Recognitioncited by 95
  7. 2021Interventional Video Grounding with Dual Contrastive Learningcited by 94
  8. 2022Debiased Learning from Naturally Imbalanced Pseudo-Labelscited by 85
  9. 2021Explainability of deep vision-based autonomous driving systems: Review and challengescited by 80
  10. 2021Counterfactual Invariance to Spurious Correlations: Why and How to Pass Stress Testscited by 77
  11. 2022Cross-Modal Causal Relational Reasoning for Event-Level Visual Question Answeringcited by 57
  12. 2021Using latent space regression to analyze and leverage compositionality in GANscited by 55
38 more
  1. 2022Mechanistic Mode Connectivitycited by 50
  2. 2021Learning Disentangled Representations in the Imaging Domaincited by 46
  3. 2022A Causal Lens for Controllable Text Generationcited by 45
  4. 2022From Statistical to Causal Learningcited by 40
  5. 2021Counterfactual State Explanations for Reinforcement Learning Agents via Generative Deep Learningcited by 39
  6. 2021Transporting Causal Mechanisms for Unsupervised Domain Adaptationcited by 38
  7. 2022Invariant and Transportable Representations for Anti-Causal Domain Shiftscited by 37
  8. 2024Exploring Diffusion Time-steps for Unsupervised Representation Learningcited by 36
  9. 2024Emergence of Hidden Capabilities: Exploring Learning Dynamics in Concept Spacecited by 35
  10. 2022Invariant Feature Learning for Generalized Long-Tailed Classificationcited by 34
  11. 2023Measuring axiomatic soundness of counterfactual image modelscited by 34
  12. 2024LTGC: Long-tail Recognition via Leveraging LLMs-driven Generated Contentcited by 33
  13. 2022Symmetry-Based Representations for Artificial and Biological General Intelligencecited by 26
  14. 2022ComCLIP: Training-Free Compositional Image and Text Matchingcited by 26
  15. 2021Learning Domain Invariant Representations for Generalizable Person Re-Identificationcited by 22
  16. 2022CPL: Counterfactual Prompt Learning for Vision and Language Modelscited by 18
  17. 2021The Blessings of Unlabeled Background in Untrimmed Videoscited by 16
  18. 2023Visual Causal Scene Refinement for Video Question Answeringcited by 16
  19. 2023Targeted Reduction of Causal Modelscited by 15
  20. 2021Improving Weakly-supervised Object Localization via Causal Interventioncited by 14
  21. 2021Towards Unbiased Visual Emotion Recognition via Causal Interventioncited by 11
  22. 2025Visual Jenga: Discovering Object Dependencies via Counterfactual Inpaintingcited by 11
  23. 2023Emerging Synergies in Causality and Deep Generative Models: A Surveycited by 10
  24. 2024Interpreting Low-level Vision Models with Causal Effect Mapscited by 8
  25. 2021VDSM: Unsupervised Video Disentanglement with State-Space Modeling and Deep Mixtures of Expertscited by 6
  26. 2022Intelligent Computing: The Latest Advances, Challenges and Futurecited by 6
  27. 2024Towards the Reusability and Compositionality of Causal Representationscited by 5
  28. 2026Steer Like the LLM: Activation Steering that Mimics Promptingcited by 5
  29. 2023VCD: Visual Causality Discovery for Cross-Modal Question Reasoningcited by 2
  30. 2023AIRIVA: A Deep Generative Model of Adaptive Immune Repertoirescited by 2
  31. 2021Hierarchical Variational Autoencoder for Visual Counterfactualscited by 1
  32. 2022Explanatory causal effects for model agnostic explanationscited by 1
  33. 2023SCADI: Self-supervised Causal Disentanglement in Latent Variable Modelscited by 1
  34. 2024Bayesian Unsupervised Disentanglement of Anatomy and Geometry for Deep Groupwise Image Registrationcited by 1
  35. 2021Recovering Barab'asi-Albert Parameters of Graphs through Disentanglementcited by 0
  36. 2022Attribute Graphs Underlying Molecular Generative Models: Path to Learning with Limited Datacited by 0
  37. 2022Nonlinear Causal Discovery via Kernel Anchor Regressioncited by 0
  38. 2023CausalAPM: Generalizable Literal Disentanglement for NLU Debiasingcited by 0

50 most cited of 51 All 51 (JSON)