mlsanityData to 7 Oct 2026

MONet: Unsupervised Scene Decomposition and Representation

arXiv 1901.11390 ↗First posted 22 Jan 2019cs.CV cs.LG stat.MLcited by 388

Not on the map. Why?

Cited by 31 in World models, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

The ability to decompose scenes in terms of abstract building blocks is crucial for general intelligence. Where those basic building blocks share meaningful properties, interactions and other regularities across scenes, such decompositions can simplify reasoning and facilitate imagination of novel scenarios. In particular, representing perceptual observations in terms of entities should improve data efficiency and transfer performance on a wide range of tasks. Thus we need models capable of discovering useful decompositions of scenes by identifying units with such regularities and representing them in a common format. To address this problem, we have developed the Multi-Object Network (MONet). In this model, a VAE is trained end-to-end together with a recurrent attention network -- in a purely unsupervised manner -- to provide attention masks around, and reconstructions of, regions of images. We show that this model is capable of learning to decompose and represent challenging 3D scenes into semantically meaningful components, such as objects and background elements.

Key facts. Cited by 388 papers in the map (data through 7 Oct 2026). First posted on arXiv 22 Jan 2019.

Builds on

No reference list yet.

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2019: 1; 2020: 0; 2021: 71; 2022: 77; 2023: 76; 2024: 64; 2025: 48; 2026: 51.

388 papers.

Sort:

By area:

  1. 2022Kubric: A scalable dataset generatorcited by 373
  2. 2024ReKep: Spatio-Temporal Reasoning of Relational Keypoint Constraints for Robotic Manipulationcited by 305
  3. 2021Conditional Object-Centric Learning from Videocited by 247
  4. 2021Localizing Objects with Self-Supervised Transformers and no Labelscited by 227
  5. 2022Bridging the Gap to Real-World Object-Centric Learningcited by 219
  6. 2021D'ya like DAGs? A Survey on Structure Learning and Causal Discoverycited by 209
  7. 2021Towards Causal Representation Learningcited by 207
  8. 2022SAVi++: Towards End-to-End Object-Centric Learning from Real-World Videoscited by 165
  9. 2022Deep Spectral Methods: A Surprisingly Strong Baseline for Unsupervised Semantic Segmentation and Localizationcited by 150
  10. 2021Self-supervised Video Object Segmentation by Motion Groupingcited by 147
  11. 2021Illiterate DALL-E Learns to Composecited by 147
  12. 2022SlotFormer: Unsupervised Visual Dynamics Simulation with Object-Centric Modelscited by 146
38 more
  1. 2021Physion: Evaluating Physical Prediction from Vision in Humans and Machinescited by 145
  2. 2022Simple Unsupervised Object-Centric Learning for Complex and Naturalistic Videoscited by 139
  3. 2022VIOLA: Imitation Learning for Vision-Based Manipulation with Object Proposal Priorscited by 135
  4. 2021Generative Adversarial Transformerscited by 120
  5. 2021GENESIS-V2: Inferring Unordered Object Representations without Iterative Refinementcited by 109
  6. 2021Unsupervised Discovery of Object Radiance Fieldscited by 100
  7. 2023Consciousness in Artificial Intelligence: Insights from the Science of Consciousnesscited by 99
  8. 2021ClevrTex: A Texture-Rich Benchmark for Unsupervised Multi-Object Segmentationcited by 93
  9. 2022Object Scene Representation Transformercited by 90
  10. 2023Object-Centric Slot Diffusioncited by 90
  11. 2023SlotDiffusion: Object-Centric Generative Modeling with Diffusion Modelscited by 85
  12. 2022Object discovery and representation networkscited by 80
  13. 2023Object-Centric Learning for Real-World Videos by Predicting Temporal Feature Similaritiescited by 78
  14. 2021Neural Production Systems: Learning Rule-Governed Visual Dynamicscited by 75
  15. 2021Unsupervised Learning of Compositional Energy Conceptscited by 71
  16. 2024Understanding the Limits of Vision Language Models Through the Lens of the Binding Problemcited by 71
  17. 2022Self-Supervised Visual Representation Learning with Semantic Groupingcited by 70
  18. 2021SIMONe: View-Invariant, Temporally-Abstracted Object Representations via Unsupervised Video Decompositioncited by 69
  19. 2022Learning Multi-Object Dynamics with Compositional Neural Radiance Fieldscited by 63
  20. 2022Segmenting Moving Objects via an Object-Centric Layered Representationcited by 63
  21. 2023Compositional Generalization from First Principlescited by 62
  22. 2022ComPhy: Compositional Physical Reasoning of Objects and Events from Videoscited by 61
  23. 2022Improving Object-centric Learning with Query Optimizationcited by 60
  24. 2023Learning Generalizable Manipulation Policies with Object-Centric 3D Representationscited by 60
  25. 2021SemanticStyleGAN: Learning Compositional Generative Priors for Controllable Image Synthesis and Editingcited by 58
  26. 2021Unsupervised Part Discovery from Contrastive Reconstructioncited by 57
  27. 2023Additive Decoders for Latent Variables Identification and Cartesian-Product Extrapolationcited by 56
  28. 2021Large-Scale Unsupervised Object Discoverycited by 54
  29. 2022Discovering Object Masks with Transformers for Unsupervised Semantic Segmentationcited by 54
  30. 2022Unsupervised Semantic Segmentation with Self-supervised Object-centric Representationscited by 54
  31. 2023Physion++: Evaluating Physical Scene Understanding that Requires Online Inference of Different Physical Propertiescited by 53
  32. 2021Attention, please! A survey of Neural Attention Models in Deep Learningcited by 50
  33. 2021Unsupervised Layered Image Decomposition into Object Prototypescited by 50
  34. 2023An Investigation into Pre-Training Object-Centric Representations for Reinforcement Learningcited by 49
  35. 2023Self-supervised Object-Centric Learning for Videoscited by 49
  36. 2023SPOT: Self-Training with Patch-Order Permutation for Object-Centric Learning with Autoregressive Transformerscited by 49
  37. 2022Visual Reinforcement Learning with Self-Supervised 3D Representationscited by 45
  38. 2023Provably Learning Object-Centric Representationscited by 45

50 most cited of 388 All 388 (JSON)