mlsanityData to 14 Oct 2026

Isolating Sources of Disentanglement in Variational Autoencoders

arXiv 1802.04942 ↗First posted 14 Feb 2018cs.LG cs.AI stat.MLcited by 167

Not on the map. Why?

Abstract (arXiv)

We decompose the evidence lower bound to show the existence of a term measuring the total correlation between latent variables. We use this to motivate our β-TCVAE (Total Correlation Variational Autoencoder), a refinement of the state-of-the-art β-VAE objective for learning disentangled representations, requiring no additional hyperparameters during training. We further propose a principled classifier-free measure of disentanglement called the mutual information gap (MIG). We perform extensive quantitative and qualitative experiments, in both restricted and non-restricted settings, and show a strong relation between total correlation and disentanglement, when the latent variables model is trained using our framework.

Key facts. Cited by 167 papers in the map (data through 14 Oct 2026). First posted on arXiv 14 Feb 2018.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2019: 2; 2020: 1; 2021: 75; 2022: 54; 2023: 20; 2024: 5; 2025: 5; 2026: 5.

167 papers.

Sort:

By area:

  1. 2021Interpretable Machine Learning: Fundamental Principles and 10 Grand Challengescited by 367
  2. 2021Self-Supervised Learning with Data Augmentations Provably Isolates Content from Stylecited by 312
  3. 2019Causal Confusion in Imitation Learningcited by 288
  4. 2021Disentanglement via Mechanism Sparsity Regularization: A New Principle for Nonlinear ICAcited by 160
  5. 2022Disentangled Representation Learningcited by 112
  6. 2021Towards Principled Disentanglement for Domain Generalizationcited by 91
  7. 2021Repurposing GANs for One-shot Semantic Part Segmentationcited by 88
  8. 2021Generalization and Robustness Implications in Object-Centric Learningcited by 81
  9. 2021Desiderata for Representation Learning: A Causal Perspectivecited by 76
  10. 2022GlanceNets: Interpretabile, Leak-proof Concept-based Modelscited by 73
  11. 2021Visual Representation Learning Does Not Generalize Strongly Within the Same Domaincited by 61
  12. 2021Semantics Disentangling for Generalized Zero-Shot Learningcited by 56
38 more
  1. 2022Diversify and Disambiguate: Learning From Underspecified Datacited by 56
  2. 2023High Fidelity Image Counterfactuals with Probabilistic Causal Modelscited by 53
  3. 2021Self-Supervised Learning Disentangled Group Representation as Featurecited by 47
  4. 2021Disentangling Hate in Online Memescited by 46
  5. 2022Synergies between Disentanglement and Sparsity: Generalization and Identifiability in Multi-Task Learningcited by 44
  6. 2022Compositional Generalization in Unsupervised Compositional Representation Learning: A Study on Disentanglement and Emergent Languagecited by 42
  7. 2021STEEX: Steering Counterfactual Explanations with Semanticscited by 40
  8. 2021Orthogonal Jacobian Regularization for Unsupervised Disentanglement in Image Generationcited by 35
  9. 2021Contrastively Disentangled Sequential Variational Autoencodercited by 34
  10. 2022Identifying Weight-Variant Latent Causal Modelscited by 34
  11. 2022Learning Causal Representations of Single Cells via Sparse Mechanism Shift Modelingcited by 34
  12. 2021An Image is Worth More Than a Thousand Words: Towards Disentanglement in the Wildcited by 30
  13. 2021Echocardiography Segmentation with Enforced Temporal Consistencycited by 30
  14. 2022Disentanglement with Biological Constraints: A Theory of Functional Cell Typescited by 30
  15. 2021Multi-Facet Clustering Variational Autoencoderscited by 28
  16. 2021Multimodal Representation Learning via Maximization of Local Mutual Informationcited by 27
  17. 2021Where and What? Examining Interpretable Disentangled Representationscited by 27
  18. 2022Symmetry-Based Representations for Artificial and Biological General Intelligencecited by 26
  19. 2021Self-Supervised Disentangled Representation Learning for Third-Person Imitation Learningcited by 23
  20. 2021Generative Adversarial Graph Convolutional Networks for Human Action Synthesiscited by 23
  21. 2022DCI-ES: An Extended Disentanglement Framework with Connections to Identifiabilitycited by 23
  22. 2023Interpretability is in the Mind of the Beholder: A Causal Framework for Human-interpretable Representation Learningcited by 22
  23. 2021I Don't Need u: Identifiable Non-Linear ICA Without Side Informationcited by 21
  24. 2022On PAC-Bayesian reconstruction guarantees for VAEscited by 20
  25. 2021Disentangling Generative Factors in Natural Language with Discrete Variational Autoencoderscited by 19
  26. 2022Learning Disentangled Representations for Counterfactual Regression via Mutual Information Minimizationcited by 19
  27. 2022On the Versatile Uses of Partial Distance Correlation in Deep Learningcited by 19
  28. 2021Lifelong Teacher-Student Network Learningcited by 18
  29. 2021Disentangled Motif-aware Graph Learning for Phrase Groundingcited by 17
  30. 2021Local Disentanglement in Variational Auto-Encoders Using Jacobian L1 Regularizationcited by 17
  31. 2022Attri-VAE: attribute-based interpretable representations of medical images with variational autoencoderscited by 17
  32. 2021Disentangling Representations of Text by Masking Transformerscited by 15
  33. 2021Group Equivariant Subsamplingcited by 15
  34. 2021Variational Attention: Propagating Domain-Specific Knowledge for Multi-Domain Learning in Crowd Countingcited by 15
  35. 2022Interpretable Molecular Graph Generation via Monotonic Constraintscited by 15
  36. 2021Disentangled Representations from Non-Disentangled Modelscited by 13
  37. 2022Learning Disentangled Representations of Negation and Uncertaintycited by 13
  38. 2022NashAE: Disentangling Representations through Adversarial Covariance Minimizationcited by 13

50 most cited of 167 All 167 (JSON)