mlsanityData to 14 Oct 2026

Weakly Supervised Disentanglement with Guarantees

arXiv 1910.09772 ↗First posted 22 Oct 2019cs.LG stat.MLcited by 82

Not on the map. Why?

Abstract (arXiv)

Learning disentangled representations that correspond to factors of variation in real-world data is critical to interpretable and human-controllable machine learning. Recently, concerns about the viability of learning disentangled representations in a purely unsupervised manner has spurred a shift toward the incorporation of weak supervision. However, there is currently no formalism that identifies when and how weak supervision will guarantee disentanglement. To address this issue, we provide a theoretical framework to assist in analyzing the disentanglement guarantees (or lack thereof) conferred by weak supervision when coupled with learning algorithms based on distribution matching. We empirically verify the guarantees and limitations of several weak supervision methods (restricted labeling, match-pairing, and rank-pairing), demonstrating the predictive power and usefulness of our theoretical framework.

Key facts. Cited by 82 papers in the map (data through 14 Oct 2026). First posted on arXiv 22 Oct 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2021: 22; 2022: 15; 2023: 21; 2024: 15; 2025: 7; 2026: 2.

82 papers.

Sort:

By area:

  1. 2021Self-Supervised Learning with Data Augmentations Provably Isolates Content from Stylecited by 312
  2. 2021Towards Causal Representation Learningcited by 209
  3. 2022Weakly supervised causal representation learningcited by 140
  4. 2022Causality Inspired Representation Learning for Domain Generalizationcited by 138
  5. 2023Nonparametric Identifiability of Causal Representations from Unknown Interventionscited by 106
  6. 2021Desiderata for Representation Learning: A Causal Perspectivecited by 76
  7. 2022GlanceNets: Interpretabile, Leak-proof Concept-based Modelscited by 73
  8. 2023Not All Neuro-Symbolic Concepts Are Created Equal: Analysis and Mitigation of Reasoning Shortcutscited by 63
  9. 2021Visual Representation Learning Does Not Generalize Strongly Within the Same Domaincited by 61
  10. 2023Identifiability Results for Multimodal Contrastive Learningcited by 61
  11. 2023Score-based Causal Representation Learning with Interventionscited by 49
  12. 2023Provable Compositional Generalization for Object-Centric Learningcited by 41
38 more
  1. 2023Disentanglement via Latent Quantizationcited by 40
  2. 2021DISSECT: Disentangled Simultaneous Explanations via Concept Traversalscited by 37
  3. 2023Measuring axiomatic soundness of counterfactual image modelscited by 34
  4. 2023General Identifiability and Achievability for Causal Representation Learningcited by 34
  5. 2021Properties from Mechanisms: An Equivariance Perspective on Identifiable Representation Learningcited by 30
  6. 2023Conditional Mutual Information for Disentangled Representations in Reinforcement Learningcited by 27
  7. 2021Interactive Disentanglement: Learning Concepts by Interacting with their Prototype Representationscited by 24
  8. 2022DCI-ES: An Extended Disentanglement Framework with Connections to Identifiabilitycited by 23
  9. 2021On Causally Disentangled Representationscited by 22
  10. 2023Interpretability is in the Mind of the Beholder: A Causal Framework for Human-interpretable Representation Learningcited by 22
  11. 2021Learning disentangled representations via product manifold projectioncited by 20
  12. 2022Temporal Disentanglement of Representations for Improved Generalisation in Reinforcement Learningcited by 19
  13. 2022On the Versatile Uses of Partial Distance Correlation in Deep Learningcited by 19
  14. 2024Analyzing (In)Abilities of SAEs via Formal Languagescited by 17
  15. 2023Improving Compositional Generalization Using Iterated Learning and Simplicial Embeddingscited by 15
  16. 2021Addressing the Topological Defects of Disentanglement via Distributed Operatorscited by 14
  17. 2022Towards efficient representation identification in supervised learningcited by 14
  18. 2024Neural Concept Bindercited by 14
  19. 2022Controlling Directions Orthogonal to a Classifiercited by 13
  20. 2022Learning Disentangled Representations of Negation and Uncertaintycited by 13
  21. 2023Towards a Unified Framework of Contrastive Learning for Disentangled Representationscited by 13
  22. 2026WS-GRPO: Weakly-Supervised Group-Relative Policy Optimization for Rollout-Efficient Reasoningcited by 12
  23. 2024Counterfactual Generative Modeling with Variational Causal Inferencecited by 11
  24. 2021Exploring the Latent Space of Autoencoders with Interventional Assayscited by 10
  25. 2023Simple Disentanglement of Style and Content in Visual Representationscited by 10
  26. 2022Provable concept learning for interpretable predictions using variational autoencoderscited by 9
  27. 2021Disentanglement and Generalization Under Correlation Shiftscited by 8
  28. 2024A survey on Concept-based Approaches For Model Improvementcited by 8
  29. 2024Double InfoGAN for Contrastive Analysiscited by 7
  30. 2021VDSM: Unsupervised Video Disentanglement with State-Space Modeling and Deep Mixtures of Expertscited by 6
  31. 2023Disentangled Variational Autoencoder for Emotion Recognition in Conversationscited by 5
  32. 2021Controllable Data Augmentation Through Deep Relightingcited by 4
  33. 2024Towards Controllable Time Series Generationcited by 4
  34. 2024Disentangling Representations through Multi-task Learningcited by 4
  35. 2021Video Reenactment as Inductive Bias for Content-Motion Disentanglementcited by 3
  36. 2021Learning ABCs: Approximate Bijective Correspondence for isolating factors of variation with weak supervisioncited by 3
  37. 2023A Category-theoretical Meta-analysis of Definitions of Disentanglementcited by 3
  38. 2023Conditionally Invariant Representation Learning for Disentangling Cellular Heterogeneitycited by 3

50 most cited of 82 All 82 (JSON)