---
title: "Physion — paper by Bear et al. (2021) — summary and ideas"
description: "Vision algorithms do well on many tasks, but it is unclear whether they grasp how physical scenes… 70 references, cited by 13 papers here."
authors: ["Daniel M. Bear", "Elias Wang", "Damian Mrowca", "Felix J. Binder", "Hsiao-Yu Fish Tung", "R. T. Pramod", "Cameron Holdaway", "Sirui Tao", "Kevin Smith", "Fan-Yun Sun", "Li Fei-Fei", "Nancy Kanwisher", "Joshua B. Tenenbaum", "Daniel L. K. Yamins", "Judith E. Fan"]
year: 2021
venue: "arXiv:2106.08261"
source: "https://arxiv.org/abs/2106.08261"
doi: "10.48550/arxiv.2106.08261"
arxiv: "2106.08261"
licence: "arXiv — non-exclusive licence to distribute"
licence_url: "http://arxiv.org/licenses/nonexclusive-distrib/1.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/bear-2021-physion-evaluating-physical-prediction-vision/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/bear-2021-physion-evaluating-physical-prediction-vision/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2021](https://md.research.elamur.ai/free/en/papers/2021/page/6/) / Physion: Evaluating Physical Prediction from Vision in Humans and Machines

# Physion: Evaluating Physical Prediction from Vision in Humans and Machines

**Authors:** Daniel M. Bear, Elias Wang, Damian Mrowca, Felix J. Binder, Hsiao-Yu Fish Tung, R. T. Pramod, Cameron Holdaway, Sirui Tao, Kevin Smith, Fan-Yun Sun, [Li Fei-Fei](https://md.research.elamur.ai/free/en/authors/li-fei-fei/), Nancy Kanwisher, [Joshua B. Tenenbaum](https://md.research.elamur.ai/free/en/authors/joshua-b-tenenbaum/), Daniel L. K. Yamins and Judith E. Fan · **Year:** 2021 · **Venue:** arXiv:2106.08261 · **Licence:** [arXiv — non-exclusive licence to distribute](http://arxiv.org/licenses/nonexclusive-distrib/1.0/)

**Links:** [Publisher's page](https://arxiv.org/abs/2106.08261) · [PDF](https://arxiv.org/pdf/2106.08261) · [DOI 10.48550/arxiv.2106.08261](https://doi.org/10.48550/arxiv.2106.08261) · [arXiv 2106.08261](https://arxiv.org/abs/2106.08261) · [OpenAlex](https://openalex.org/W3169703926)

**Citations:** 10 citations (OpenAlex, 2026-10-14) · [13](https://md.research.elamur.ai/free/en/papers/bear-2021-physion-evaluating-physical-prediction-vision/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
Vision algorithms do well on many tasks, but it is unclear whether they grasp how physical scenes change over time. Earlier evaluations either required predicting every pixel or trajectory, or covered only one narrow phenomenon, such as falling block towers. The authors created Physion, a benchmark of realistic 3D simulations built in the ThreeDWorld (TDW) simulator. It contains eight scenario types and uses a single object contact prediction (OCP) task: will two cued objects touch as the scene unfolds? Because OCP does not depend on model architecture, very different models can be compared under the same conditions. The authors also measured human predictions on the same videos. Many model families were tested, including pixel-level predictors, object-centric vision models and graph neural networks given particle or physical-state input. No vision model reached human-level performance. Object-centric vision models generally beat the others but still trailed people. Particle-based graph models did much better and made more human-like predictions. The authors conclude that extracting physical scene representations from raw images is the main bottleneck.

## Key points
- **Eight scenarios:** Dominoes, Support, Collide, Contain, Drop, Link, Roll and Drape. Together they cover rigid and soft-body collisions, stability, rolling, sliding, projectile motion and cloth behaviour. All were rendered with TDW, which runs on Unity3D.
- **Evaluation design:** OCP is scored through several readout protocols and three training regimes: training on all scenarios, on a single scenario ("only"), or on all but the test scenario ("all-but"). Particle-input models were evaluated only with the observed+simulated readout.
- **Training regimes:** For vision models, switching between regimes changed results little. The "only" regime gave a small gain (b=0.21, p\<10⁻⁴), and "all-but" caused no significant drop (p=0.057). However, models trained on one scenario transferred poorly to most others, which shows that diverse training data matters.
- **Main finding:** Explicit, object- or particle-level physical representations predict better and more like humans. The authors connect this to the debate over symbolic representations in AI.
- **Release:** All data, code for generating extra training data, and benchmarking code are public, so new models can be evaluated reproducibly.

*Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.*

## Institutions

- [Stanford University](https://md.research.elamur.ai/free/en/universities/stanford-university/) — University, Stanford, United States · 567 works · 335,189 citations
- [Massachusetts Institute of Technology](https://md.research.elamur.ai/free/en/universities/massachusetts-institute-of-technology/) — University, Cambridge, United States · 274 works · 82,013 citations
- [University of California San Diego](https://md.research.elamur.ai/free/en/universities/university-of-california-san-diego/) — University, San Diego, United States · 101 works · 54,571 citations
- [Indian Institute of Science Bangalore](https://md.research.elamur.ai/free/en/universities/indian-institute-of-science-bangalore/) — University, Bengaluru, India · 16 works · 1,462 citations
- [McGovern Institute for Brain Research](https://md.research.elamur.ai/free/en/labs/mcgovern-institute-for-brain-research/) — Research institute, Cambridge, United States · 4 works · 305 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 19,503 works · [👁️ Computer Vision and Pattern Recognition](https://md.research.elamur.ai/free/en/shelves/computer-science/computer-vision-and-pattern-recognition/) · 3,621 works · [⚙️ Engineering](https://md.research.elamur.ai/free/en/shelves/engineering/) · 206 works · [⚙️ Control and Systems Engineering](https://md.research.elamur.ai/free/en/shelves/engineering/control-and-systems-engineering/) · 72 works

## Cite

Cite as: Daniel M. Bear, Elias Wang, Damian Mrowca, Felix J. Binder, Hsiao-Yu Fish Tung, R. T. Pramod, Cameron Holdaway, Sirui Tao, Kevin Smith, Fan-Yun Sun, Li Fei-Fei, Nancy Kanwisher, Joshua B. Tenenbaum, Daniel L. K. Yamins and Judith E. Fan. 2021. Physion: Evaluating Physical Prediction from Vision in Humans and Machines. arXiv:2106.08261.

The paper itself, not this page. BibTeX from the library's record of the work.

```bibtex
@misc{bear2021physion,
  title = {{Physion: Evaluating Physical Prediction from Vision in Humans and Machines}},
  author = {Daniel M. Bear, Elias Wang, Damian Mrowca, Felix J. Binder, Hsiao-Yu Fish Tung, R. T. Pramod, Cameron Holdaway, Sirui Tao, Kevin Smith, Fan-Yun Sun, Li Fei-Fei, Nancy Kanwisher, Joshua B. Tenenbaum, Daniel L. K. Yamins and Judith E. Fan},
  year = {2021},
  doi = {10.48550/arxiv.2106.08261},
  eprint = {2106.08261},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.2106.08261}
}
```

- [← previous Multi-tasking Dialogue Comprehension with Discourse Parsing](https://md.research.elamur.ai/free/en/papers/he-2021-multi-tasking-dialogue-comprehension-discourse/)
- [next → A Large-Scale English Multi-Label Twitter Dataset for Cyberbullying and Online Abuse Detection](https://md.research.elamur.ai/free/en/papers/salawu-2021-large-scale-english-multi-label/)
- [↑ back to papers of 2021](https://md.research.elamur.ai/free/en/papers/2021/page/6/)

## References

70 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. P. Agrawal, A. V. Nair, P. Abbeel, J. Malik, and S. Levine. Learning to poke by poking: Experiential learning of intuitive physics. In Advances in Neural Information Processing Systems, pages 5074–5082, 2016. [find](https://scholar.google.com/scholar?q=P%2E+Agrawal%2C+A%2E+V%2E+Nair%2C+P%2E+Abbeel%2C+J%2E+Malik%2C+and+S%2E+Levine%2E+Learning+to+poke+by+poking%3A+Experiential+learning+of+intuitive+physics%2E+In+Advances+in+Neural+Information+Processing+Systems%2C+pages+5074%E2%80%93508)
2. A. Ajay, M. Bauza, J. Wu, N. Fazeli, J. B. Tenenbaum, A. Rodriguez, and L. P. Kaelbling. Combining physical simulators and object-based networks for control. In International Conference on Robotics and Automation, 2019. [find](https://scholar.google.com/scholar?q=A%2E+Ajay%2C+M%2E+Bauza%2C+J%2E+Wu%2C+N%2E+Fazeli%2C+J%2E+B%2E+Tenenbaum%2C+A%2E+Rodriguez%2C+and+L%2E+P%2E+Kaelbling%2E+Combining+physical+simulators+and+object%2Dbased+networks+for+control%2E+In+International+Conference+on+Robotics+an)
3. K. R. Allen, K. A. Smith, and J. B. Tenenbaum. Rapid trial-and-error learning with simulation supports flexible tool use and physical reasoning. Proceedings of the National Academy of Sciences, 117(47): 29302–29310, 2020. ISSN 0027-8424, 1091-6490. doi: 10.1073/pnas.1912341117. — [doi:10.1073/pnas.1912341117](https://doi.org/10.1073/pnas.1912341117) [find](https://scholar.google.com/scholar?q=K%2E+R%2E+Allen%2C+K%2E+A%2E+Smith%2C+and+J%2E+B%2E+Tenenbaum%2E+Rapid+trial%2Dand%2Derror+learning+with+simulation+supports+flexible+tool+use+and+physical+reasoning%2E+Proceedings+of+the+National+Academy+of+Sciences%2C+117%2847)
4. R. Baillargeon, E. S. Spelke, and S. Wasserman. Object permanence in five-month-old infants. Cognition, 20(3):191–208, 1985. [find](https://scholar.google.com/scholar?q=R%2E+Baillargeon%2C+E%2E+S%2E+Spelke%2C+and+S%2E+Wasserman%2E+Object+permanence+in+five%2Dmonth%2Dold+infants%2E+Cognition%2C+20%283%29%3A191%E2%80%93208%2C+1985%2E)
5. R. Baillargeon, J. Li, Y. Gertner, and D. Wu. How do infants reason about physical events? Wiley-Blackwell, 2011. [find](https://scholar.google.com/scholar?q=R%2E+Baillargeon%2C+J%2E+Li%2C+Y%2E+Gertner%2C+and+D%2E+Wu%2E+How+do+infants+reason+about+physical+events%3F+Wiley%2DBlackwell%2C+2011%2E)
6. A. Bakhtin, L. van der Maaten, J. Johnson, L. Gustafson, and R. Girshick. [Phyre: A new benchmark for physical reasoning](https://md.research.elamur.ai/free/en/papers/bakhtin-2019-phyre-new-benchmark-physical-reasoning/). Advances in Neural Information Processing Systems, 32:5082–5093, 2019.
7. C. J. Bates, P. W. Battaglia, I. Yildirim, and J. B. Tenenbaum. Humans predict liquid dynamics using probabilistic simulation. In CogSci, 2015. [find](https://scholar.google.com/scholar?q=C%2E+J%2E+Bates%2C+P%2E+W%2E+Battaglia%2C+I%2E+Yildirim%2C+and+J%2E+B%2E+Tenenbaum%2E+Humans+predict+liquid+dynamics+using+probabilistic+simulation%2E+In+CogSci%2C+2015%2E)
8. C. J. Bates, I. Yildirim, J. B. Tenenbaum, and P. W. Battaglia. Modeling human intuitions about liquid flow with particle-based simulation. PLoS Computational Biology, 2019. [find](https://scholar.google.com/scholar?q=C%2E+J%2E+Bates%2C+I%2E+Yildirim%2C+J%2E+B%2E+Tenenbaum%2C+and+P%2E+W%2E+Battaglia%2E+Modeling+human+intuitions+about+liquid+flow+with+particle%2Dbased+simulation%2E+PLoS+Computational+Biology%2C+2019%2E)
9. P. Battaglia, R. Pascanu, M. Lai, D. J. Rezende, et al. Interaction networks for learning about objects, relations and physics. In Advances in Neural Information Processing Systems, pages 4502–4510, 2016. [find](https://scholar.google.com/scholar?q=P%2E+Battaglia%2C+R%2E+Pascanu%2C+M%2E+Lai%2C+D%2E+J%2E+Rezende%2C+et+al%2E+Interaction+networks+for+learning+about+objects%2C+relations+and+physics%2E+In+Advances+in+Neural+Information+Processing+Systems%2C+pages+4502%E2%80%934510%2C+2)
10. P. W. Battaglia, J. B. Hamrick, and J. B. Tenenbaum. Simulation as an engine of physical scene understanding. Proceedings of the National Academy of Sciences, 110(45):18327–18332, 2013. [find](https://scholar.google.com/scholar?q=P%2E+W%2E+Battaglia%2C+J%2E+B%2E+Hamrick%2C+and+J%2E+B%2E+Tenenbaum%2E+Simulation+as+an+engine+of+physical+scene+understanding%2E+Proceedings+of+the+National+Academy+of+Sciences%2C+110%2845%29%3A18327%E2%80%9318332%2C+2013%2E)
11. P. W. Battaglia, J. B. Hamrick, V. Bapst, A. Sanchez-Gonzalez, V. Zambaldi, M. Malinowski, A. Tacchetti, D. Raposo, A. Santoro, R. Faulkner, et al. [Relational inductive biases, deep learning, and graph networks](https://md.research.elamur.ai/free/en/papers/battaglia-2018-relational-inductive-biases-deep-learning/). arXiv preprint arXiv:1806.01261, 2018.
12. D. M. Bear, C. Fan, D. Mrowca, Y. Li, S. Alter, A. Nayebi, J. Schwartz, L. Fei-Fei, J. Wu, J. B. Tenenbaum, et al. Learning physical graph representations from visual scenes. In Advances in Neural Information Processing Systems, 2020. [find](https://scholar.google.com/scholar?q=D%2E+M%2E+Bear%2C+C%2E+Fan%2C+D%2E+Mrowca%2C+Y%2E+Li%2C+S%2E+Alter%2C+A%2E+Nayebi%2C+J%2E+Schwartz%2C+L%2E+Fei%2DFei%2C+J%2E+Wu%2C+J%2E+B%2E+Tenenbaum%2C+et+al%2E+Learning+physical+graph+representations+from+visual+scenes%2E+In+Advances+in+Neural+Inf)
13. W. Bi, A. D. Shah, K. W. Wong, B. Scholl, and I. Yildirim. Perception of soft materials relies on physics-based object representations: Behavioral and computational evidence. bioRxiv, 2021. [find](https://scholar.google.com/scholar?q=W%2E+Bi%2C+A%2E+D%2E+Shah%2C+K%2E+W%2E+Wong%2C+B%2E+Scholl%2C+and+I%2E+Yildirim%2E+Perception+of+soft+materials+relies+on+physics%2Dbased+object+representations%3A+Behavioral+and+computational+evidence%2E+bioRxiv%2C+2021%2E)
14. C. P. Burgess, L. Matthey, N. Watters, R. Kabra, I. Higgins, M. Botvinick, and A. Lerchner. Monet: Unsupervised scene decomposition and representation. arXiv preprint arXiv:1901.11390, 2019. — [arXiv:1901.11390](https://arxiv.org/abs/1901.11390) [find](https://scholar.google.com/scholar?q=C%2E+P%2E+Burgess%2C+L%2E+Matthey%2C+N%2E+Watters%2C+R%2E+Kabra%2C+I%2E+Higgins%2C+M%2E+Botvinick%2C+and+A%2E+Lerchner%2E+Monet%3A+Unsupervised+scene+decomposition+and+representation%2E+arXiv+preprint+arXiv%3A1901%2E11390%2C+2019%2E+%E2%80%94+%5BarXiv%3A)
15. S. Carey and F. Xu. Infants’ knowledge of objects: Beyond object files and object tracking. Cognition, 80 (1-2):179–213, 2001. [find](https://scholar.google.com/scholar?q=S%2E+Carey+and+F%2E+Xu%2E+Infants%E2%80%99+knowledge+of+objects%3A+Beyond+object+files+and+object+tracking%2E+Cognition%2C+80+%281%2D2%29%3A179%E2%80%93213%2C+2001%2E)
16. M. B. Chang, T. Ullman, A. Torralba, and J. B. Tenenbaum. A compositional object-based approach to learning physical dynamics. In International Conference on Learning Representations, 2017. [find](https://scholar.google.com/scholar?q=M%2E+B%2E+Chang%2C+T%2E+Ullman%2C+A%2E+Torralba%2C+and+J%2E+B%2E+Tenenbaum%2E+A+compositional+object%2Dbased+approach+to+learning+physical+dynamics%2E+In+International+Conference+on+Learning+Representations%2C+2017%2E)
17. S. Dasari, F. Ebert, S. Tian, S. Nair, B. Bucher, K. Schmeckpeper, S. Singh, S. Levine, and C. Finn. Robonet: Large-scale multi-robot learning. arXiv preprint arXiv:1910.11215, 2019. — [arXiv:1910.11215](https://arxiv.org/abs/1910.11215) [find](https://scholar.google.com/scholar?q=S%2E+Dasari%2C+F%2E+Ebert%2C+S%2E+Tian%2C+S%2E+Nair%2C+B%2E+Bucher%2C+K%2E+Schmeckpeper%2C+S%2E+Singh%2C+S%2E+Levine%2C+and+C%2E+Finn%2E+Robonet%3A+Large%2Dscale+multi%2Drobot+learning%2E+arXiv+preprint+arXiv%3A1910%2E11215%2C+2019%2E+%E2%80%94+%5BarXiv%3A1910%2E112)
18. E. Denton and R. Fergus. Stochastic video generation with a learned prior. In International Conference on Machine Learning, pages 1174–1183. PMLR, 2018. [find](https://scholar.google.com/scholar?q=E%2E+Denton+and+R%2E+Fergus%2E+Stochastic+video+generation+with+a+learned+prior%2E+In+International+Conference+on+Machine+Learning%2C+pages+1174%E2%80%931183%2E+PMLR%2C+2018%2E)
19. D. Ding, F. Hill, A. Santoro, and M. Botvinick. Object-based attention for spatio-temporal reasoning: Outperforming neuro-symbolic models with flexible distributed architectures. arXiv preprint arXiv:2012.08508, 2020. — [arXiv:2012.08508](https://arxiv.org/abs/2012.08508) [find](https://scholar.google.com/scholar?q=D%2E+Ding%2C+F%2E+Hill%2C+A%2E+Santoro%2C+and+M%2E+Botvinick%2E+Object%2Dbased+attention+for+spatio%2Dtemporal+reasoning%3A+Outperforming+neuro%2Dsymbolic+models+with+flexible+distributed+architectures%2E+arXiv+preprint+arXiv%3A)
20. F. Ebert, S. Dasari, A. X. Lee, S. Levine, and C. Finn. Robustness via retrying: Closed-loop robotic manipulation with self-supervised learning. In Conference on Robot Learning, pages 983–993. PMLR, 2018. [find](https://scholar.google.com/scholar?q=F%2E+Ebert%2C+S%2E+Dasari%2C+A%2E+X%2E+Lee%2C+S%2E+Levine%2C+and+C%2E+Finn%2E+Robustness+via+retrying%3A+Closed%2Dloop+robotic+manipulation+with+self%2Dsupervised+learning%2E+In+Conference+on+Robot+Learning%2C+pages+983%E2%80%93993%2E+PMLR%2C+2)
21. C. Finn, I. Goodfellow, and S. Levine. Unsupervised learning for physical interaction through video prediction. In Advances in Neural Information Processing Systems, pages 64–72, 2016. [find](https://scholar.google.com/scholar?q=C%2E+Finn%2C+I%2E+Goodfellow%2C+and+S%2E+Levine%2E+Unsupervised+learning+for+physical+interaction+through+video+prediction%2E+In+Advances+in+Neural+Information+Processing+Systems%2C+pages+64%E2%80%9372%2C+2016%2E)
22. M. A. Fischler and R. C. Bolles. Random sample consensus: A paradigm for model fitting with applications to image analysis and automated cartography. Commun. ACM, 24(6):381–395, June 1981. ISSN 0001-0782. doi: 10.1145/358669.358692. URL https://doi.org/10.1145/358669.358692. [↗](https://doi.org/10.1145/358669.358692)
23. K. Fragkiadaki, P. Agrawal, S. Levine, and J. Malik. Learning visual predictive models of physics for playing billiards. In International Conference on Learning Representations, 2016. [find](https://scholar.google.com/scholar?q=K%2E+Fragkiadaki%2C+P%2E+Agrawal%2C+S%2E+Levine%2C+and+J%2E+Malik%2E+Learning+visual+predictive+models+of+physics+for+playing+billiards%2E+In+International+Conference+on+Learning+Representations%2C+2016%2E)
24. C. Gan, J. Schwartz, S. Alter, M. Schrimpf, J. Traer, J. De Freitas, J. Kubilius, A. Bhandwaldar, N. Haber, M. Sano, K. Kim, E. Wang, D. Mrowca, M. Lingelbach, A. Curtis, K. Feigelis, D. M. Bear, D. Gutfreund, D. Cox, J. J. DiCarlo, J. McDermott, J. B. Tenenbaum, and D. L. K. Yamins. ThreeDWorld: A… — [arXiv:2007.04954](https://arxiv.org/abs/2007.04954) [find](https://scholar.google.com/scholar?q=C%2E+Gan%2C+J%2E+Schwartz%2C+S%2E+Alter%2C+M%2E+Schrimpf%2C+J%2E+Traer%2C+J%2E+De+Freitas%2C+J%2E+Kubilius%2C+A%2E+Bhandwaldar%2C+N%2E+Haber%2C+M%2E+Sano%2C+K%2E+Kim%2C+E%2E+Wang%2C+D%2E+Mrowca%2C+M%2E+Lingelbach%2C+A%2E+Curtis%2C+K%2E+Feigelis%2C+D%2E+M%2E+Bear%2C+D%2E+G)
25. T. Gebru, J. Morgenstern, B. Vecchione, J. W. Vaughan, H. Wallach, H. Daumé III, and K. Crawford. Datasheets for datasets. arXiv preprint arXiv:1803.09010, 2018. [find](https://scholar.google.com/scholar?q=T%2E+Gebru%2C+J%2E+Morgenstern%2C+B%2E+Vecchione%2C+J%2E+W%2E+Vaughan%2C+H%2E+Wallach%2C+H%2E+Daum%C3%A9+III%2C+and+K%2E+Crawford%2E+Datasheets+for+datasets%2E+arXiv+preprint+arXiv%3A1803%2E09010%2C+2018%2E)
26. R. Geirhos, J.-H. Jacobsen, C. Michaelis, R. Zemel, W. Brendel, M. Bethge, and F. A. Wichmann. Shortcut learning in deep neural networks. Nature Machine Intelligence, 2(11):665–673, 2020. [find](https://scholar.google.com/scholar?q=R%2E+Geirhos%2C+J%2E%2DH%2E+Jacobsen%2C+C%2E+Michaelis%2C+R%2E+Zemel%2C+W%2E+Brendel%2C+M%2E+Bethge%2C+and+F%2E+A%2E+Wichmann%2E+Shortcut+learning+in+deep+neural+networks%2E+Nature+Machine+Intelligence%2C+2%2811%29%3A665%E2%80%93673%2C+2020%2E)
27. R. Girdhar, L. Gustafson, A. Adcock, and L. van der Maaten. Forward prediction for physical reasoning. arXiv preprint arXiv:2006.10734, 2020. — [arXiv:2006.10734](https://arxiv.org/abs/2006.10734) [find](https://scholar.google.com/scholar?q=R%2E+Girdhar%2C+L%2E+Gustafson%2C+A%2E+Adcock%2C+and+L%2E+van+der+Maaten%2E+Forward+prediction+for+physical+reasoning%2E+arXiv+preprint+arXiv%3A2006%2E10734%2C+2020%2E+%E2%80%94+%5BarXiv%3A2006%2E10734%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2006%2E10734%29)
28. R. Girshick. Fast r-cnn. In International Conference on Computer Vision, pages 1440–1448, 2015. [find](https://scholar.google.com/scholar?q=R%2E+Girshick%2E+Fast+r%2Dcnn%2E+In+International+Conference+on+Computer+Vision%2C+pages+1440%E2%80%931448%2C+2015%2E)
29. O. Groth, F. B. Fuchs, I. Posner, and A. Vedaldi. Shapestacks: Learning vision-based physical intuition for generalised object stacking. In European Conference on Computer Vision, pages 702–717, 2018. [find](https://scholar.google.com/scholar?q=O%2E+Groth%2C+F%2E+B%2E+Fuchs%2C+I%2E+Posner%2C+and+A%2E+Vedaldi%2E+Shapestacks%3A+Learning+vision%2Dbased+physical+intuition+for+generalised+object+stacking%2E+In+European+Conference+on+Computer+Vision%2C+pages+702%E2%80%93717%2C+2018%2E)
30. N. Haber, D. Mrowca, L. Fei-Fei, and D. L. Yamins. Learning to play with intrinsically-motivated self-aware agents. In Advances in Neural Information Processing Systems, 2018. [find](https://scholar.google.com/scholar?q=N%2E+Haber%2C+D%2E+Mrowca%2C+L%2E+Fei%2DFei%2C+and+D%2E+L%2E+Yamins%2E+Learning+to+play+with+intrinsically%2Dmotivated+self%2Daware+agents%2E+In+Advances+in+Neural+Information+Processing+Systems%2C+2018%2E)
31. H. Hecht and M. Bertamini. Understanding projectile acceleration. Journal of Experimental Psychology: Human Perception and Performance, 26(2):730–746, 2000. ISSN 1939-1277, 0096-1523. doi: 10.1037/0096-1523.26.2.730. — [doi:10.1037/0096-1523.26.2.730](https://doi.org/10.1037/0096-1523.26.2.730) [find](https://scholar.google.com/scholar?q=H%2E+Hecht+and+M%2E+Bertamini%2E+Understanding+projectile+acceleration%2E+Journal+of+Experimental+Psychology%3A+Human+Perception+and+Performance%2C+26%282%29%3A730%E2%80%93746%2C+2000%2E+ISSN+1939%2D1277%2C+0096%2D1523%2E+doi%3A+10%2E1037%2F009)
32. M. Janner, S. Levine, W. T. Freeman, J. B. Tenenbaum, C. Finn, and J. Wu. Reasoning about physical interactions with object-oriented prediction and planning. In International Conference on Learning Representations, 2019. [find](https://scholar.google.com/scholar?q=M%2E+Janner%2C+S%2E+Levine%2C+W%2E+T%2E+Freeman%2C+J%2E+B%2E+Tenenbaum%2C+C%2E+Finn%2C+and+J%2E+Wu%2E+Reasoning+about+physical+interactions+with+object%2Doriented+prediction+and+planning%2E+In+International+Conference+on+Learning+Re)
33. T. Kipf, E. van der Pol, and M. Welling. Contrastive learning of structured world models. In International Conference on Learning Representations, 2020. [find](https://scholar.google.com/scholar?q=T%2E+Kipf%2C+E%2E+van+der+Pol%2C+and+M%2E+Welling%2E+Contrastive+learning+of+structured+world+models%2E+In+International+Conference+on+Learning+Representations%2C+2020%2E)
34. A. X. Lee, R. Zhang, F. Ebert, P. Abbeel, C. Finn, and S. Levine. Stochastic adversarial video prediction. arXiv preprint arXiv:1804.01523, 2018. — [arXiv:1804.01523](https://arxiv.org/abs/1804.01523) [find](https://scholar.google.com/scholar?q=A%2E+X%2E+Lee%2C+R%2E+Zhang%2C+F%2E+Ebert%2C+P%2E+Abbeel%2C+C%2E+Finn%2C+and+S%2E+Levine%2E+Stochastic+adversarial+video+prediction%2E+arXiv+preprint+arXiv%3A1804%2E01523%2C+2018%2E+%E2%80%94+%5BarXiv%3A1804%2E01523%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F1804%2E01523%29)
35. A. Lerer, S. Gross, and R. Fergus. Learning physical intuition of block towers by example. In International Conference on Machine Learning, 2016. [find](https://scholar.google.com/scholar?q=A%2E+Lerer%2C+S%2E+Gross%2C+and+R%2E+Fergus%2E+Learning+physical+intuition+of+block+towers+by+example%2E+In+International+Conference+on+Machine+Learning%2C+2016%2E)
36. W. Li, S. Azimi, A. Leonardis, and M. Fritz. To fall or not to fall: A visual approach to physical stability prediction. arXiv preprint arXiv:1604.00066, 2016. — [arXiv:1604.00066](https://arxiv.org/abs/1604.00066) [find](https://scholar.google.com/scholar?q=W%2E+Li%2C+S%2E+Azimi%2C+A%2E+Leonardis%2C+and+M%2E+Fritz%2E+To+fall+or+not+to+fall%3A+A+visual+approach+to+physical+stability+prediction%2E+arXiv+preprint+arXiv%3A1604%2E00066%2C+2016%2E+%E2%80%94+%5BarXiv%3A1604%2E00066%5D%28https%3A%2F%2Farxiv%2Eorg%2Fa)
37. Y. Li, J. Wu, R. Tedrake, J. B. Tenenbaum, and A. Torralba. Learning particle dynamics for manipulating rigid bodies, deformable objects, and fluids. In International Conference on Learning Representations, 2019. [find](https://scholar.google.com/scholar?q=Y%2E+Li%2C+J%2E+Wu%2C+R%2E+Tedrake%2C+J%2E+B%2E+Tenenbaum%2C+and+A%2E+Torralba%2E+Learning+particle+dynamics+for+manipulating+rigid+bodies%2C+deformable+objects%2C+and+fluids%2E+In+International+Conference+on+Learning+Representa)
38. Y. Li, T. Lin, K. Yi, D. Bear, D. Yamins, J. Wu, J. Tenenbaum, and A. Torralba. Visual grounding of learned physical models. In International Conference on Machine Learning, pages 5927–5936. PMLR, 2020. [find](https://scholar.google.com/scholar?q=Y%2E+Li%2C+T%2E+Lin%2C+K%2E+Yi%2C+D%2E+Bear%2C+D%2E+Yamins%2C+J%2E+Wu%2C+J%2E+Tenenbaum%2C+and+A%2E+Torralba%2E+Visual+grounding+of+learned+physical+models%2E+In+International+Conference+on+Machine+Learning%2C+pages+5927%E2%80%935936%2E+PMLR%2C+202)
39. A. Mandlekar, Y. Zhu, A. Garg, J. Booher, M. Spero, A. Tung, J. Gao, J. Emmons, A. Gupta, E. Orbay, et al. Roboturk: A crowdsourcing platform for robotic skill learning through imitation. In Conference on Robot Learning, pages 879–893. PMLR, 2018. [find](https://scholar.google.com/scholar?q=A%2E+Mandlekar%2C+Y%2E+Zhu%2C+A%2E+Garg%2C+J%2E+Booher%2C+M%2E+Spero%2C+A%2E+Tung%2C+J%2E+Gao%2C+J%2E+Emmons%2C+A%2E+Gupta%2C+E%2E+Orbay%2C+et+al%2E+Roboturk%3A+A+crowdsourcing+platform+for+robotic+skill+learning+through+imitation%2E+In+Conferenc)
40. R. Mottaghi, H. Bagherinezhad, M. Rastegari, and A. Farhadi. Newtonian scene understanding: Unfolding the dynamics of objects in static images. In Conference on Computer Vision and Pattern Recognition, pages 3521–3529, 2016. [find](https://scholar.google.com/scholar?q=R%2E+Mottaghi%2C+H%2E+Bagherinezhad%2C+M%2E+Rastegari%2C+and+A%2E+Farhadi%2E+Newtonian+scene+understanding%3A+Unfolding+the+dynamics+of+objects+in+static+images%2E+In+Conference+on+Computer+Vision+and+Pattern+Recognition)
41. R. Mottaghi, M. Rastegari, A. Gupta, and A. Farhadi. “what happens if...” learning to predict the effect of forces in images. In European Conference on Computer Vision, pages 269–285. Springer, 2016. [find](https://scholar.google.com/scholar?q=R%2E+Mottaghi%2C+M%2E+Rastegari%2C+A%2E+Gupta%2C+and+A%2E+Farhadi%2E+%E2%80%9Cwhat+happens+if%2E%2E%2E%E2%80%9D+learning+to+predict+the+effect+of+forces+in+images%2E+In+European+Conference+on+Computer+Vision%2C+pages+269%E2%80%93285%2E+Springer%2C+2016%2E)
42. D. Mrowca, C. Zhuang, E. Wang, N. Haber, L. F. Fei-Fei, J. Tenenbaum, and D. L. Yamins. Flexible neural representation for physics prediction. In Advances in Neural Information Processing Systems, pages 8813–8824, 2018. [find](https://scholar.google.com/scholar?q=D%2E+Mrowca%2C+C%2E+Zhuang%2C+E%2E+Wang%2C+N%2E+Haber%2C+L%2E+F%2E+Fei%2DFei%2C+J%2E+Tenenbaum%2C+and+D%2E+L%2E+Yamins%2E+Flexible+neural+representation+for+physics+prediction%2E+In+Advances+in+Neural+Information+Processing+Systems%2C+pag)
43. S. Nair, S. Savarese, and C. Finn. Goal-aware prediction: Learning to model what matters. In International Conference on Machine Learning, pages 7207–7219. PMLR, 2020. [find](https://scholar.google.com/scholar?q=S%2E+Nair%2C+S%2E+Savarese%2C+and+C%2E+Finn%2E+Goal%2Daware+prediction%3A+Learning+to+model+what+matters%2E+In+International+Conference+on+Machine+Learning%2C+pages+7207%E2%80%937219%2E+PMLR%2C+2020%2E)
44. R. S. Nickerson and M. J. Adams. Long-term memory for a common object. Cognitive Psychology, 11(3): 287–307, 1979. ISSN 00100285. doi: 10.1016/0010-0285(79)90013-6. — [doi:10.1016/0010-0285(79)90013-6](https://doi.org/10.1016/0010-0285%2879%2990013-6) [find](https://scholar.google.com/scholar?q=R%2E+S%2E+Nickerson+and+M%2E+J%2E+Adams%2E+Long%2Dterm+memory+for+a+common+object%2E+Cognitive+Psychology%2C+11%283%29%3A+287%E2%80%93307%2C+1979%2E+ISSN+00100285%2E+doi%3A+10%2E1016%2F0010%2D0285%2879%2990013%2D6%2E+%E2%80%94+%5Bdoi%3A10%2E1016%2F0010%2D0285%2879%2990013%2D6)
45. A. Paszke, S. Gross, F. Massa, A. Lerer, J. Bradbury, G. Chanan, T. Killeen, Z. Lin, N. Gimelshein, L. Antiga, A. Desmaison, A. Kopf, E. Yang, Z. DeVito, M. Raison, A. Tejani, S. Chilamkurthy, B. Steiner, L. Fang, J. Bai, and S. Chintala. Pytorch: An imperative style, high-performance deep learning… — [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://md.research.elamur.ai/free/en/papers/paszke-2019-pytorch-imperative-style-high-performance/)
46. L. Piloto, A. Weinstein, A. Ahuja, M. Mirza, G. Wayne, D. Amos, C.-c. Hung, and M. Botvinick. Probing physics knowledge using tools from developmental psychology. arXiv preprint arXiv:1804.01128, 2018. — [arXiv:1804.01128](https://arxiv.org/abs/1804.01128) [find](https://scholar.google.com/scholar?q=L%2E+Piloto%2C+A%2E+Weinstein%2C+A%2E+Ahuja%2C+M%2E+Mirza%2C+G%2E+Wayne%2C+D%2E+Amos%2C+C%2E%2Dc%2E+Hung%2C+and+M%2E+Botvinick%2E+Probing+physics+knowledge+using+tools+from+developmental+psychology%2E+arXiv+preprint+arXiv%3A1804%2E01128%2C+2018)
47. H. Qi, X. Wang, D. Pathak, Y. Ma, and J. Malik. Learning long-term visual dynamics with region proposal interaction networks. In International Conference on Learning Representations, 2021. [find](https://scholar.google.com/scholar?q=H%2E+Qi%2C+X%2E+Wang%2C+D%2E+Pathak%2C+Y%2E+Ma%2C+and+J%2E+Malik%2E+Learning+long%2Dterm+visual+dynamics+with+region+proposal+interaction+networks%2E+In+International+Conference+on+Learning+Representations%2C+2021%2E)
48. R. Rajalingham, E. B. Issa, P. Bashivan, K. Kar, K. Schmidt, and J. J. DiCarlo. Large-scale, high-resolution comparison of the core visual object recognition behavior of humans, monkeys, and state-of-the-art deep artificial neural networks. Journal of Neuroscience, 38(33):7255–7269, 2018. [find](https://scholar.google.com/scholar?q=R%2E+Rajalingham%2C+E%2E+B%2E+Issa%2C+P%2E+Bashivan%2C+K%2E+Kar%2C+K%2E+Schmidt%2C+and+J%2E+J%2E+DiCarlo%2E+Large%2Dscale%2C+high%2Dresolution+comparison+of+the+core+visual+object+recognition+behavior+of+humans%2C+monkeys%2C+and+state%2Dof%2D)
49. R. Riochet, M. Y. Castro, M. Bernard, A. Lerer, R. Fergus, V. Izard, and E. Dupoux. Intphys: A framework and benchmark for visual intuitive physics reasoning. arXiv preprint arXiv:1803.07616, 2018. [find](https://scholar.google.com/scholar?q=R%2E+Riochet%2C+M%2E+Y%2E+Castro%2C+M%2E+Bernard%2C+A%2E+Lerer%2C+R%2E+Fergus%2C+V%2E+Izard%2C+and+E%2E+Dupoux%2E+Intphys%3A+A+framework+and+benchmark+for+visual+intuitive+physics+reasoning%2E+arXiv+preprint+arXiv%3A1803%2E07616%2C+2018%2E)
50. R. Riochet, J. Sivic, I. Laptev, and E. Dupoux. Occlusion resistant learning of intuitive physics from videos. arXiv:2005.00069 [cs, eess], 2020. — [arXiv:2005.00069](https://arxiv.org/abs/2005.00069) [find](https://scholar.google.com/scholar?q=R%2E+Riochet%2C+J%2E+Sivic%2C+I%2E+Laptev%2C+and+E%2E+Dupoux%2E+Occlusion+resistant+learning+of+intuitive+physics+from+videos%2E+arXiv%3A2005%2E00069+%5Bcs%2C+eess%5D%2C+2020%2E+%E2%80%94+%5BarXiv%3A2005%2E00069%5D%28https%3A%2F%2Farxiv%2Eorg%2Fabs%2F2005%2E00069%29)
51. A. N. Sanborn, V. K. Mansinghka, and T. L. Griffiths. Reconciling intuitive physics and Newtonian mechanics for colliding objects. Psychological Review, 120(2):411–437, 2013. ISSN 0033-295X. doi: http://dx.doi.org/10.1037/a0031912. — [doi:10.1037/a0031912](https://doi.org/10.1037/a0031912) [↗](http://dx.doi.org/10.1037/a0031912)
52. A. Sanchez-Gonzalez, N. Heess, J. T. Springenberg, J. Merel, M. Riedmiller, R. Hadsell, and P. W. Battaglia. Graph networks as learnable physics engines for inference and control. In International Conference on Machine Learning, 2018. [find](https://scholar.google.com/scholar?q=A%2E+Sanchez%2DGonzalez%2C+N%2E+Heess%2C+J%2E+T%2E+Springenberg%2C+J%2E+Merel%2C+M%2E+Riedmiller%2C+R%2E+Hadsell%2C+and+P%2E+W%2E+Battaglia%2E+Graph+networks+as+learnable+physics+engines+for+inference+and+control%2E+In+International+Con)
53. A. Sanchez-Gonzalez, J. Godwin, T. Pfaff, R. Ying, J. Leskovec, and P. Battaglia. Learning to simulate complex physics with graph networks. In International Conference on Machine Learning, pages 8459–8468. PMLR, 2020. [find](https://scholar.google.com/scholar?q=A%2E+Sanchez%2DGonzalez%2C+J%2E+Godwin%2C+T%2E+Pfaff%2C+R%2E+Ying%2C+J%2E+Leskovec%2C+and+P%2E+Battaglia%2E+Learning+to+simulate+complex+physics+with+graph+networks%2E+In+International+Conference+on+Machine+Learning%2C+pages+8459%E2%80%93)
54. K. Schmeckpeper, A. Xie, O. Rybkin, S. Tian, K. Daniilidis, S. Levine, and C. Finn. Learning predictive models from observation and interaction. In European Conference on Computer Vision, 2020. [find](https://scholar.google.com/scholar?q=K%2E+Schmeckpeper%2C+A%2E+Xie%2C+O%2E+Rybkin%2C+S%2E+Tian%2C+K%2E+Daniilidis%2C+S%2E+Levine%2C+and+C%2E+Finn%2E+Learning+predictive+models+from+observation+and+interaction%2E+In+European+Conference+on+Computer+Vision%2C+2020%2E)
55. D. J. Simons and M. S. Ambinder. Change Blindness: Theory and Consequences. Current Directions in Psychological Science, 14(1):44–48, 2005. ISSN 0963-7214, 1467-8721. doi: 10.1111/j.0963-7214.2005.00332.x. — [doi:10.1111/j.0963-7214.2005.00332.x](https://doi.org/10.1111/j.0963-7214.2005.00332.x) [find](https://scholar.google.com/scholar?q=D%2E+J%2E+Simons+and+M%2E+S%2E+Ambinder%2E+Change+Blindness%3A+Theory+and+Consequences%2E+Current+Directions+in+Psychological+Science%2C+14%281%29%3A44%E2%80%9348%2C+2005%2E+ISSN+0963%2D7214%2C+1467%2D8721%2E+doi%3A+10%2E1111%2Fj%2E0963%2D7214%2E2005%2E003)
56. K. Simonyan and A. Zisserman. [Very deep convolutional networks for large-scale image recognition](https://md.research.elamur.ai/free/en/papers/simonyan-2014-very-deep-convolutional-networks-large/). In International Conference on Learning Representations, 2015.
57. K. Smith, L. Mei, S. Yao, J. Wu, E. Spelke, J. Tenenbaum, and T. Ullman. Modeling expectation violation in intuitive physics with coarse probabilistic object representations. Advances in Neural Information Processing Systems, 32:8985–8995, 2019. [find](https://scholar.google.com/scholar?q=K%2E+Smith%2C+L%2E+Mei%2C+S%2E+Yao%2C+J%2E+Wu%2C+E%2E+Spelke%2C+J%2E+Tenenbaum%2C+and+T%2E+Ullman%2E+Modeling+expectation+violation+in+intuitive+physics+with+coarse+probabilistic+object+representations%2E+Advances+in+Neural+Inform)
58. K. A. Smith and E. Vul. Sources of Uncertainty in Intuitive Physics. Topics in Cognitive Science, 5(1): 185–199, 2013. ISSN 17568757. doi: 10.1111/tops.12009. — [doi:10.1111/tops.12009](https://doi.org/10.1111/tops.12009) [find](https://scholar.google.com/scholar?q=K%2E+A%2E+Smith+and+E%2E+Vul%2E+Sources+of+Uncertainty+in+Intuitive+Physics%2E+Topics+in+Cognitive+Science%2C+5%281%29%3A+185%E2%80%93199%2C+2013%2E+ISSN+17568757%2E+doi%3A+10%2E1111%2Ftops%2E12009%2E+%E2%80%94+%5Bdoi%3A10%2E1111%2Ftops%2E12009%5D%28https%3A%2F%2Fdoi%2Eor)
59. K. A. Smith, P. W. Battaglia, and E. Vul. Different Physical Intuitions Exist Between Tasks, Not Domains. Computational Brain & Behavior, 1(2):101–118, 2018. ISSN 2522-0861, 2522-087X. doi: 10.1007/s42113-018-0007-3. — [doi:10.1007/s42113-018-0007-3](https://doi.org/10.1007/s42113-018-0007-3) [find](https://scholar.google.com/scholar?q=K%2E+A%2E+Smith%2C+P%2E+W%2E+Battaglia%2C+and+E%2E+Vul%2E+Different+Physical+Intuitions+Exist+Between+Tasks%2C+Not+Domains%2E+Computational+Brain+%26+Behavior%2C+1%282%29%3A101%E2%80%93118%2C+2018%2E+ISSN+2522%2D0861%2C+2522%2D087X%2E+doi%3A+10%2E1007%2Fs4)
60. E. S. Spelke. Principles of object perception. Cognitive science, 14(1):29–56, 1990. [find](https://scholar.google.com/scholar?q=E%2E+S%2E+Spelke%2E+Principles+of+object+perception%2E+Cognitive+science%2C+14%281%29%3A29%E2%80%9356%2C+1990%2E)
61. A. Tacchetti, H. F. Song, P. A. M. Mediano, V. F. Zambaldi, J. Kramár, N. C. Rabinowitz, T. Graepel, M. Botvinick, and P. W. Battaglia. Relational forward models for multi-agent learning. In International Conference on Learning Representations, 2019. [find](https://scholar.google.com/scholar?q=A%2E+Tacchetti%2C+H%2E+F%2E+Song%2C+P%2E+A%2E+M%2E+Mediano%2C+V%2E+F%2E+Zambaldi%2C+J%2E+Kram%C3%A1r%2C+N%2E+C%2E+Rabinowitz%2C+T%2E+Graepel%2C+M%2E+Botvinick%2C+and+P%2E+W%2E+Battaglia%2E+Relational+forward+models+for+multi%2Dagent+learning%2E+In+Internati)
62. H. Touvron, M. Cord, M. Douze, F. Massa, A. Sablayrolles, and H. Jégou. [Training data-efficient image transformers & distillation through attention](https://md.research.elamur.ai/free/en/papers/touvron-2020-training-data-efficient-image-transformers/). arXiv preprint arXiv:2012.12877, 2020.
63. T. D. Ullman, E. Spelke, P. Battaglia, and J. B. Tenenbaum. Mind Games: Game Engines as an Architecture for Intuitive Physics. Trends in Cognitive Sciences, 21(9):649–665, 2017. ISSN 1364-6613. doi: 10.1016/j.tics.2017.05.012. [find](https://scholar.google.com/scholar?q=T%2E+D%2E+Ullman%2C+E%2E+Spelke%2C+P%2E+Battaglia%2C+and+J%2E+B%2E+Tenenbaum%2E+Mind+Games%3A+Game+Engines+as+an+Architecture+for+Intuitive+Physics%2E+Trends+in+Cognitive+Sciences%2C+21%289%29%3A649%E2%80%93665%2C+2017%2E+ISSN+1364%2D6613%2E+doi%3A+1)
64. R. Veerapaneni, J. D. Co-Reyes, M. Chang, M. Janner, C. Finn, J. Wu, J. Tenenbaum, and S. Levine. Entity abstraction in visual model-based reinforcement learning. In Conference on Robot Learning, pages 1439–1456, 2020. [find](https://scholar.google.com/scholar?q=R%2E+Veerapaneni%2C+J%2E+D%2E+Co%2DReyes%2C+M%2E+Chang%2C+M%2E+Janner%2C+C%2E+Finn%2C+J%2E+Wu%2C+J%2E+Tenenbaum%2C+and+S%2E+Levine%2E+Entity+abstraction+in+visual+model%2Dbased+reinforcement+learning%2E+In+Conference+on+Robot+Learning%2C+page)
65. R. Villegas, A. Pathak, H. Kannan, D. Erhan, Q. V. Le, and H. Lee. High fidelity video prediction with large stochastic recurrent neural networks. In Advances in Neural Information Processing Systems, 2019. [find](https://scholar.google.com/scholar?q=R%2E+Villegas%2C+A%2E+Pathak%2C+H%2E+Kannan%2C+D%2E+Erhan%2C+Q%2E+V%2E+Le%2C+and+H%2E+Lee%2E+High+fidelity+video+prediction+with+large+stochastic+recurrent+neural+networks%2E+In+Advances+in+Neural+Information+Processing+Systems%2C)
66. C. Vondrick, H. Pirsiavash, and A. Torralba. Anticipating visual representations from unlabeled video. In Conference on Computer Vision and Pattern Recognition, pages 98–106, 2016. [find](https://scholar.google.com/scholar?q=C%2E+Vondrick%2C+H%2E+Pirsiavash%2C+and+A%2E+Torralba%2E+Anticipating+visual+representations+from+unlabeled+video%2E+In+Conference+on+Computer+Vision+and+Pattern+Recognition%2C+pages+98%E2%80%93106%2C+2016%2E)
67. N. Watters, L. Matthey, M. Bosnjak, C. P. Burgess, and A. Lerchner. Cobra: Data-efficient model-based rl through unsupervised object discovery and curiosity-driven exploration. arXiv preprint arXiv:1905.09275, 2019. — [arXiv:1905.09275](https://arxiv.org/abs/1905.09275) [find](https://scholar.google.com/scholar?q=N%2E+Watters%2C+L%2E+Matthey%2C+M%2E+Bosnjak%2C+C%2E+P%2E+Burgess%2C+and+A%2E+Lerchner%2E+Cobra%3A+Data%2Defficient+model%2Dbased+rl+through+unsupervised+object+discovery+and+curiosity%2Ddriven+exploration%2E+arXiv+preprint+arXiv%3A19)
68. B. Wu, S. Nair, R. Martin-Martin, L. Fei-Fei, and C. Finn. Greedy hierarchical variational autoencoders for large-scale video prediction. arXiv preprint arXiv:2103.04174, 2021. [find](https://scholar.google.com/scholar?q=B%2E+Wu%2C+S%2E+Nair%2C+R%2E+Martin%2DMartin%2C+L%2E+Fei%2DFei%2C+and+C%2E+Finn%2E+Greedy+hierarchical+variational+autoencoders+for+large%2Dscale+video+prediction%2E+arXiv+preprint+arXiv%3A2103%2E04174%2C+2021%2E)
69. Y. Ye, M. Singh, A. Gupta, and S. Tulsiani. Compositional video prediction. In International Conference on Computer Vision, pages 10353–10362, 2019. [find](https://scholar.google.com/scholar?q=Y%2E+Ye%2C+M%2E+Singh%2C+A%2E+Gupta%2C+and+S%2E+Tulsiani%2E+Compositional+video+prediction%2E+In+International+Conference+on+Computer+Vision%2C+pages+10353%E2%80%9310362%2C+2019%2E)
70. R. Zhang, J. Wu, C. Zhang, W. T. Freeman, and J. B. Tenenbaum. A comparative evaluation of approximate probabilistic simulation and deep neural networks as accounts of human physical scene understanding. In CogSci, 2016. [find](https://scholar.google.com/scholar?q=R%2E+Zhang%2C+J%2E+Wu%2C+C%2E+Zhang%2C+W%2E+T%2E+Freeman%2C+and+J%2E+B%2E+Tenenbaum%2E+A+comparative+evaluation+of+approximate+probabilistic+simulation+and+deep+neural+networks+as+accounts+of+human+physical+scene+understandi)

*Updated 3 November 2026*