On the Identifiability of Quantized Factors
Not on the map. Why?
Abstract (arXiv)
Disentanglement aims to recover meaningful latent ground-truth factors from the observed distribution solely, and is formalized through the theory of identifiability. The identifiability of independent latent factors is proven to be impossible in the unsupervised i.i.d. setting under a general nonlinear map from factors to observations. In this work, however, we demonstrate that it is possible to recover quantized latent factors under a generic nonlinear diffeomorphism. We only assume that the latent factors have independent discontinuities in their density, without requiring the factors to be statistically independent. We introduce this novel form of identifiability, termed quantized factor identifiability, and provide a comprehensive proof of the recovery of the quantized factors.
Key facts. Cited by 3 papers in the map (data through 7 Oct 2026). Builds on 29 of its 50 references that are in the map. First posted on arXiv 28 Jun 2023.
Builds on
29 of 50 references.
References · 29
- 2017Neural Discrete Representation Learningcited by 4,650
- 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 929
- 2019Variational Autoencoders and Nonlinear ICA: A Unifying Frameworkcited by 494
- 2018Nonlinear ICA Using Auxiliary Variables and Generalized Contrastive Learningcited by 278
- 2016Unsupervised Feature Extraction by Time-Contrastive Learning and Nonlinear ICAcited by 270
- 2020Weakly-Supervised Disentanglement Without Compromisescited by 237
- 2021Disentanglement via Mechanism Sparsity Regularization: A New Principle for Nonlinear ICAcited by 158
- 2022Weakly supervised causal representation learningcited by 140
- 2022Interventional Causal Representation Learningcited by 140
- 2022CITRIS: Causal Identifiability from Temporal Intervened Sequencescited by 132
- 2020Towards Nonlinear Disentanglement in Natural Data with Temporal Sparse Codingcited by 121
- 2020ICE-BeeM: Identifiable Conditional Energy-Based Deep Models Based on Nonlinear ICAcited by 102
17 more
- 2021Independent mechanism analysis, a new concept?cited by 96
- 2020On Disentangled Representations Learned From Correlated Datacited by 94
- 2022On the Identifiability of Nonlinear ICA: Sparsity and Beyondcited by 84
- 2016Discovering Causal Signals in Imagescited by 79
- 2021Desiderata for Representation Learning: A Causal Perspectivecited by 76
- 2022Identifiability of deep generative models without auxiliary informationcited by 70
- 2021Disentangling Identifiable Features from Noisy Data with Structured Nonlinear ICAcited by 64
- 2022Weakly Supervised Representation Learning with Sparse Perturbationscited by 64
- 2020On the Transfer of Disentangled Representations in Realistic Settingscited by 63
- 2021Identifiable Deep Generative Models via Sparse Decodingcited by 54
- 2022Function Classes for Identifiable Nonlinear Independent Component Analysiscited by 53
- 2023Provably Learning Object-Centric Representationscited by 45
- 2023Disentanglement via Latent Quantizationcited by 40
- 2022Disentanglement of Correlated Factors via Hausdorff Factorized Supportcited by 37
- 2022Towards efficient representation identification in supervised learningcited by 14
- 2022Learning Latent Causal Dynamicscited by 13
- 2021Binary Independent Component Analysis: A Non-stationarity-based Approachcited by 4
Built on by
Citing papers per year: 2025: 3.
3 papers.