mlsanityData to 7 Oct 2026

DCI-ES: An Extended Disentanglement Framework with Connections to Identifiability

arXiv 2210.00364 ↗First posted 1 Oct 2022cs.LG cs.AI cs.CVcited by 23

Not on the map. Why?

Abstract (arXiv)

In representation learning, a common approach is to seek representations which disentangle the underlying factors of variation. Eastwood & Williams (2018) proposed three metrics for quantifying the quality of such disentangled representations: disentanglement (D), completeness (C) and informativeness (I). In this work, we first connect this DCI framework to two common notions of linear and nonlinear identifiability, thereby establishing a formal link between disentanglement and the closely-related field of independent component analysis. We then propose an extended DCI-ES framework with two new measures of representation quality - explicitness (E) and size (S) - and point out how D and C can be computed for black-box predictors. Our main idea is that the functional capacity required to use a representation is an important but thus-far neglected aspect of representation quality, which we quantify using explicitness or ease-of-use (E). We illustrate the relevance of our extensions on the MPI3D and Cars3D datasets.

Key facts. Cited by 23 papers in the map (data through 7 Oct 2026). Builds on 28 of its 43 references that are in the map. First posted on arXiv 1 Oct 2022.

Builds on

28 of 43 references.

References · 28

  1. 2015Deep Residual Learning for Image Recognitioncited by 51,809
  2. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  3. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  4. 2017A Unified Approach to Interpreting Model Predictionscited by 6,488
  5. 2020Scaling Laws for Neural Language Modelscited by 6,411
  6. 2013Auto-Encoding Variational Bayescited by 3,980
  7. 2018The Lottery Ticket Hypothesis: Finding Sparse, Trainable Neural Networkscited by 2,250
  8. 2018Reconciling modern machine learning practice and the bias-variance trade-offcited by 919
  9. 2016InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Netscited by 858
  10. 2019Deep Double Descent: Where Bigger Models and More Data Hurtcited by 667
  11. 2018Disentangling by Factorisingcited by 614
  12. 2019Fantastic Generalization Measures and Where to Find Themcited by 507
16 more
  1. 2019Variational Autoencoders and Nonlinear ICA: A Unifying Frameworkcited by 494
  2. 2018Nonlinear ICA Using Auxiliary Variables and Generalized Contrastive Learningcited by 278
  3. 2018Isolating Sources of Disentanglement in Variational Autoencoderscited by 167
  4. 2019Deep ReLU Networks Have Surprisingly Few Activation Patternscited by 160
  5. 2015Deep Convolutional Inverse Graphics Networkcited by 121
  6. 2018Robustly Disentangled Causal Mechanisms: Validating Deep Representations for Interventional Robustnesscited by 108
  7. 2018Learning Deep Disentangled Embeddings with the F-Statistic Losscited by 101
  8. 2019On the Transfer of Inductive Bias from Simulation to the Real World: a New Disentanglement Datasetcited by 100
  9. 2019Weakly Supervised Disentanglement with Guaranteescited by 82
  10. 2020In Search of Robust Measures of Generalizationcited by 79
  11. 2021The Shape of Learning Curves: a Reviewcited by 67
  12. 2020A Sober Look at the Unsupervised Learning of Disentangled Representations and their Evaluationcited by 52
  13. 2020Rethinking Parameter Counting in Deep Models: Effective Dimensionality Revisitedcited by 50
  14. 2020Are wider nets better given the same number of parameters?cited by 27
  15. 2020Evaluating representations by the complexity of learning low-loss predictorscited by 16
  16. 2021The Role of Pretrained Representations for the OOD Generalization of Reinforcement Learning Agentscited by 12

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 8; 2024: 6; 2025: 5; 2026: 4.

23 papers.

Sort:

By area:

  1. 2023Nonparametric Identifiability of Causal Representations from Unknown Interventionscited by 105
  2. 2023Concept Algebra for (Score-Based) Text-Controlled Generative Modelscited by 59
  3. 2023Disentanglement via Latent Quantizationcited by 40
  4. 2023Interpretability is in the Mind of the Beholder: A Causal Framework for Human-interpretable Representation Learningcited by 22
  5. 2023From Identifiable Causal Representations to Controllable Counterfactual Generation: A Survey on Causal Generative Modelingcited by 21
  6. 2024Causal Diffusion Autoencoders: Toward Counterfactual Generation via Diffusion Probabilistic Modelscited by 17
  7. 2023Self-Supervised Disentanglement by Leveraging Structure in Data Augmentationscited by 16
  8. 2023Learning Causally Disentangled Representations via the Principle of Independent Causal Mechanismscited by 14
  9. 2024Exploring the Effectiveness of Object-Centric Representations in Visual Question Answering: Comparative Insights with Foundation Modelscited by 14
  10. 2024Identifiable Causal Representation Learning: Unsupervised, Multi-View, and Multi-Environmentcited by 12
  11. 2025Representational Similarity via Interpretable Visual Conceptscited by 5
  12. 2025From Isolation to Entanglement: When Do Interpretability Methods Identify and Disentangle Known Concepts?cited by 5
11 more
  1. 2024Disentangling Representations through Multi-task Learningcited by 4
  2. 2023Enriching Disentanglement: From Logical Definitions to Quantitative Metricscited by 2
  3. 2025From Pixels to Factors: Learning Independently Controllable State Variables for Reinforcement Learningcited by 2
  4. 2024Measuring Orthogonality in Representations of Generative Modelscited by 1
  5. 2025Disentanglement Analysis in Deep Latent Variable Models Matching Aggregate Posterior Distributionscited by 1
  6. 2024Selecting Interpretability Techniques for Healthcare Machine Learning modelscited by 0
  7. 2025Operationalizing Quantized Disentanglementcited by 0
  8. 2026Temporal Fusion Nexus: A task-agnostic multi-modal embedding model for clinical narratives and irregular time series in post-kidney transplant carecited by 0
  9. 2026Who Guards the Guardians? The Challenges of Evaluating Identifiability of Learned Representationscited by 0
  10. 2026A framework for analyzing concept representations in neural modelscited by 0
  11. 2026Concepts Worth Having: Refining VLM-Guided Concept Bottleneck Models with Minimal Annotationscited by 0