# Learning Multi-Object Dynamics with Compositional Neural Radiance Fields

arXiv: https://arxiv.org/abs/2202.11855 · HTML: https://mlsanity.com/paper/2202.11855/

## Key facts

- Cited by 63 papers in the map (data through 14 Oct 2026).
- Builds on 82 of its 102 references that are in the map.
- First posted on arXiv 24 Feb 2022.

## Builds on (most cited first, 30 of 82)

- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [DeepSDF: Learning Continuous Signed Distance Functions for Shape Representation](https://mlsanity.com/paper/1901.05103/) (2019)
- [Occupancy Networks: Learning 3D Reconstruction in Function Space](https://mlsanity.com/paper/1812.03828/) (2018)
- [Dream to Control: Learning Behaviors by Latent Imagination](https://mlsanity.com/paper/1912.01603/) (2019)
- [pixelNeRF: Neural Radiance Fields from One or Few Images](https://mlsanity.com/paper/2012.02190/) (2020)
- [D-NeRF: Neural Radiance Fields for Dynamic Scenes](https://mlsanity.com/paper/2011.13961/) (2020)
- [CLEVR: A Diagnostic Dataset for Compositional Language and Elementary Visual Reasoning](https://mlsanity.com/paper/1612.06890/) (2016)
- [Relational inductive biases, deep learning, and graph networks](https://mlsanity.com/paper/1806.01261/) (2018)
- [Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model](https://mlsanity.com/paper/1911.08265/) (2019)
- [Learning Latent Dynamics for Planning from Pixels](https://mlsanity.com/paper/1811.04551/) (2018)
- [Nerfies: Deformable Neural Radiance Fields](https://mlsanity.com/paper/2011.12948/) (2020)
- [Learning to Simulate Complex Physics with Graph Networks](https://mlsanity.com/paper/2002.09405/) (2020)
- [Learning Mesh-Based Simulation with Graph Networks](https://mlsanity.com/paper/2010.03409/) (2020)
- [Scene Representation Networks: Continuous 3D-Structure-Aware Neural Scene Representations](https://mlsanity.com/paper/1906.01618/) (2019)
- [PIFu: Pixel-Aligned Implicit Function for High-Resolution Clothed Human Digitization](https://mlsanity.com/paper/1905.05172/) (2019)
- [GIRAFFE: Representing Scenes as Compositional Generative Neural Feature Fields](https://mlsanity.com/paper/2011.12100/) (2020)
- [Convolutional Occupancy Networks](https://mlsanity.com/paper/2003.04618/) (2020)
- [Neural Scene Flow Fields for Space-Time View Synthesis of Dynamic Scenes](https://mlsanity.com/paper/2011.13084/) (2020)
- [KiloNeRF: Speeding up Neural Radiance Fields with Thousands of Tiny MLPs](https://mlsanity.com/paper/2103.13744/) (2021)
- [Neural Fields in Visual Computing and Beyond](https://mlsanity.com/paper/2111.11426/) (2021)
- [Interaction Networks for Learning about Objects, Relations and Physics](https://mlsanity.com/paper/1612.00222/) (2016)
- [CLEVRER: CoLlision Events for Video REpresentation and Reasoning](https://mlsanity.com/paper/1910.01442/) (2019)
- [Non-Rigid Neural Radiance Fields: Reconstruction and Novel View Synthesis of a Dynamic Scene From Monocular Video](https://mlsanity.com/paper/2012.12247/) (2020)
- [Time-Contrastive Networks: Self-Supervised Learning from Video](https://mlsanity.com/paper/1704.06888/) (2017)
- [Implicit Behavioral Cloning](https://mlsanity.com/paper/2109.00137/) (2021)
- [Space-time Neural Irradiance Fields for Free-Viewpoint Video](https://mlsanity.com/paper/2011.12950/) (2020)
- [MONet: Unsupervised Scene Decomposition and Representation](https://mlsanity.com/paper/1901.11390/) (2019)
- [Multi-Object Representation Learning with Iterative Variational Inference](https://mlsanity.com/paper/1903.00450/) (2019)
- [Neural Scene Graphs for Dynamic Scenes](https://mlsanity.com/paper/2011.10379/) (2020)
- [Embed to Control: A Locally Linear Latent Dynamics Model for Control from Raw Images](https://mlsanity.com/paper/1506.07365/) (2015)

## Built on by (most cited first, 30 of 63)

- [Nerfstudio: A Modular Framework for Neural Radiance Field Development](https://mlsanity.com/paper/2302.04264/) (2023)
- [Text2Motion: From Natural Language Instructions to Feasible Plans](https://mlsanity.com/paper/2303.12153/) (2023)
- [InterDiff: Generating 3D Human-Object Interactions with Physics-Informed Diffusion](https://mlsanity.com/paper/2308.16905/) (2023)
- [CLIP-Fields: Weakly Supervised Semantic Fields for Robotic Memory](https://mlsanity.com/paper/2210.05663/) (2022)
- [CG-HOI: Contact-Guided 3D Human-Object Interaction Generation](https://mlsanity.com/paper/2311.16097/) (2023)
- [Reconstruction and Simulation of Elastic Objects with Spring-Mass 3D Gaussians](https://mlsanity.com/paper/2403.09434/) (2024)
- [PhysTwin: Physics-Informed Reconstruction and Simulation of Deformable Objects from Videos](https://mlsanity.com/paper/2503.17973/) (2025)
- [PointWorld: Scaling 3D World Models for In-The-Wild Robotic Manipulation](https://mlsanity.com/paper/2601.03782/) (2026)
- [NeRF in the Palm of Your Hand: Corrective Augmentation for Robotics via Novel-View Synthesis](https://mlsanity.com/paper/2301.08556/) (2023)
- [Robot See Robot Do: Imitating Articulated Object Manipulation with Monocular 4D Reconstruction](https://mlsanity.com/paper/2409.18121/) (2024)
- [SE(3)-Equivariant Relational Rearrangement with Neural Descriptor Fields](https://mlsanity.com/paper/2211.09786/) (2022)
- [Renderable Neural Radiance Map for Visual Navigation](https://mlsanity.com/paper/2303.00304/) (2023)
- [Visuo-Tactile Transformers for Manipulation](https://mlsanity.com/paper/2210.00121/) (2022)
- [Differentiable Physics Simulation of Dynamics-Augmented Neural Objects](https://mlsanity.com/paper/2210.09420/) (2022)
- [NeRF2Real: Sim2real Transfer of Vision-guided Bipedal Motion Skills using Neural Radiance Fields](https://mlsanity.com/paper/2210.04932/) (2022)
- [Reinforcement Learning with Neural Radiance Fields](https://mlsanity.com/paper/2206.01634/) (2022)
- [RUST: Latent Neural Scene Representations from Unposed Imagery](https://mlsanity.com/paper/2211.14306/) (2022)
- [Curious Exploration via Structured World Models Yields Zero-Shot Object Manipulation](https://mlsanity.com/paper/2206.11403/) (2022)
- [Recent Trends in 3D Reconstruction of General Non-Rigid Scenes](https://mlsanity.com/paper/2403.15064/) (2024)
- [Entity-Centric Reinforcement Learning for Object Manipulation from Pixels](https://mlsanity.com/paper/2404.01220/) (2024)
- [Policy Architectures for Compositional Generalization in Control](https://mlsanity.com/paper/2203.05960/) (2022)
- [Dynamic-Resolution Model Learning for Object Pile Manipulation](https://mlsanity.com/paper/2306.16700/) (2023)
- [Neural Fields in Robotics: A Survey](https://mlsanity.com/paper/2410.20220/) (2024)
- [Learning 3D Particle-based Simulators from RGB-D Videos](https://mlsanity.com/paper/2312.05359/) (2023)
- [Dream2Real: Zero-Shot 3D Object Rearrangement with Vision-Language Models](https://mlsanity.com/paper/2312.04533/) (2023)
- [3D-IntPhys: Towards More Generalized 3D-grounded Visual Intuitive Physics under Challenging Scenes](https://mlsanity.com/paper/2304.11470/) (2023)
- [Learning rigid-body simulators over implicit shapes for large-scale scenes and vision](https://mlsanity.com/paper/2405.14045/) (2024)
- [Modeling the Real World with High-Density Visual Particle Dynamics](https://mlsanity.com/paper/2406.19800/) (2024)
- [DyST: Towards Dynamic Neural Scene Representations on Real-World Videos](https://mlsanity.com/paper/2310.06020/) (2023)
- [CompGS: Unleashing 2D Compositionality for Compositional Text-to-3D via Dynamically Optimizing 3D Gaussians](https://mlsanity.com/paper/2410.20723/) (2024)

## Abstract (arXiv)

We present a method to learn compositional multi-object dynamics models from image observations based on implicit object encoders, Neural Radiance Fields (NeRFs), and graph neural networks. NeRFs have become a popular choice for representing scenes due to their strong 3D prior. However, most NeRF approaches are trained on a single scene, representing the whole scene with a global model, making generalization to novel scenes, containing different numbers of objects, challenging. Instead, we present a compositional, object-centric auto-encoder framework that maps multiple views of the scene to a set of latent vectors representing each object separately. The latent vectors parameterize individual NeRFs from which the scene can be reconstructed. Based on those latent vectors, we train a graph neural network dynamics model in the latent space to achieve compositionality for dynamics prediction. A key feature of our approach is that the latent vectors are forced to encode 3D information through the NeRF decoder, which enables us to incorporate structural priors in learning the dynamics models, making long-term predictions more stable compared to several baselines. Simulated and real world experiments show that our method can model and learn the dynamics of compositional scenes including rigid and deformable objects. Video: https://dannydriess.github.io/compnerfdyn/