mlsanityData to 7 Oct 2026

Boxhead: A Dataset for Learning Hierarchical Representations

arXiv 2110.03628 ↗First posted 7 Oct 2021cs.LG cs.CV stat.MLcited by 1

Not on the map. Why?

Abstract (arXiv)

Disentanglement is hypothesized to be beneficial towards a number of downstream tasks. However, a common assumption in learning disentangled representations is that the data generative factors are statistically independent. As current methods are almost solely evaluated on toy datasets where this ideal assumption holds, we investigate their performance in hierarchical settings, a relevant feature of real-world data. In this work, we introduce Boxhead, a dataset with hierarchically structured ground-truth generative factors. We use this novel dataset to evaluate the performance of state-of-the-art autoencoder-based disentanglement models and observe that hierarchical models generally outperform single-layer VAEs in terms of disentanglement of hierarchically arranged factors.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 41 of its 56 references that are in the map. First posted on arXiv 7 Oct 2021.

Builds on

41 of 56 references.

References · 41

  1. 2014Deep Learning Face Attributes in the Wildcited by 3,679
  2. 2018Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representationscited by 929
  3. 2016InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Netscited by 858
  4. 2018Open3D: A Modern Library for 3D Data Processingcited by 717
  5. 2017Poincar'e Embeddings for Learning Hierarchical Representationscited by 674
  6. 2020NVAE: A Deep Hierarchical Variational Autoencodercited by 635
  7. 2018Disentangling by Factorisingcited by 614
  8. 2019Multi-Object Representation Learning with Iterative Variational Inferencecited by 359
  9. 2018Understanding disentangling in β-VAEcited by 332
  10. 2016Ladder Variational Autoencoderscited by 315
  11. 2020Very Deep VAEs Generalize Autoregressive Models and Can Outperform Them on Imagescited by 261
  12. 2019Causality for Machine Learningcited by 242
29 more
  1. 2017Variational Inference of Disentangled Latent Concepts from Unlabeled Observationscited by 240
  2. 2020Weakly-Supervised Disentanglement Without Compromisescited by 237
  3. 2019Recurrent Independent Mechanismscited by 219
  4. 2021Towards Causal Representation Learningcited by 207
  5. 2019Flexibly Fair Representation Learning by Disentanglementcited by 196
  6. 2018Isolating Sources of Disentanglement in Variational Autoencoderscited by 167
  7. 2019On the Fairness of Disentangled Representationscited by 144
  8. 2019Spatial Broadcast Decoder: A Simple Architecture for Learning Disentangled Representations in VAEscited by 116
  9. 2017Causal Consistency of Structural Equation Modelscited by 114
  10. 2019BIVA: A Very Deep Hierarchy of Latent Variables for Generative Modelingcited by 109
  11. 2018Robustly Disentangled Causal Mechanisms: Validating Deep Representations for Interventional Robustnesscited by 108
  12. 2019Are Disentangled Representations Helpful for Abstract Visual Reasoning?cited by 100
  13. 2020On Disentangled Representations Learned From Correlated Datacited by 94
  14. 2018Variational Autoencoders Pursue PCA Directions (by Accident)cited by 84
  15. 2019Weakly Supervised Disentanglement with Guaranteescited by 82
  16. 2021Generalization and Robustness Implications in Object-Centric Learningcited by 81
  17. 2020Learning Unsupervised Hierarchical Part Decomposition of 3D Objects from a Single RGB Imagecited by 72
  18. 2020On the Transfer of Disentangled Representations in Realistic Settingscited by 63
  19. 2020A Sober Look at the Unsupervised Learning of Disentangled Representations and their Evaluationcited by 52
  20. 2020Learning Physical Graph Representations from Visual Scenescited by 50
  21. 2018Life-Long Disentangled Representation Learning with Cross-Domain Latent Homologiescited by 43
  22. 2020Progressive Learning and Disentanglement of Hierarchical Representationscited by 30
  23. 2019Unsupervised Discovery of Parts, Structure, and Dynamicscited by 20
  24. 2019Disentangled State Space Representationscited by 14
  25. 2021Benchmarks, Algorithms, and Metrics for Hierarchical Disentanglementcited by 12
  26. 2020S2RMs: Spatially Structured Recurrent Modulescited by 10
  27. 2020Hierarchical Relational Inferencecited by 9
  28. 2021Spatial Dependency Networks: Neural Layers for Improved Generative Image Modelingcited by 5
  29. 2018SplineNets: Continuous Neural Decision Graphscited by 4

Built on by

Cited by year

Citing papers per year: 2022: 1.

1 paper.

  1. 2022Top-down inference in an early visual cortex inspired hierarchical Variational Autoencodercited by 2