Research map / Probabilistic and scientific modeling / Dynamical systems
Flexible neural representation for physics prediction
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Use a hierarchical graph instead of a fully connected graph (ablation) | HRN · Two-cube collisions on a plane | position MSE | Helped |
| Use a hierarchical graph instead of a fully connected graph (ablation) | HRN · Two-cube collisions on a plane | delta position MSE | Helped |
| Use a hierarchical graph instead of a fully connected graph (ablation) | HRN · Two-cube collisions on a plane | preserve distance MSE | Helped |
| Add the global loss to the local loss (ablation) | HRN · Two-cube collisions on a plane | position MSE | Helped |
| Add the global loss to the local loss (ablation) | HRN · Two-cube collisions on a plane | delta position MSE | Helped |
| Add the preservation loss (ablation) | HRN · Two-cube collisions on a plane | preserve distance MSE | Helped |
| Add the global delta position loss (ablation) | HRN · Two-cube collisions on a plane | position MSE | Helped |
| Use two input time steps instead of one (ablation) | HRN · Two-cube collisions on a plane | position MSE | Helped |
| Use separate force, collision, and history modules with separate weights (ablation) | HRN · Two-cube collisions on a plane | position MSE | Helped |
| Use hierarchical graph convolution rather than omitting the hierarchical propagation mechanism (ablation) | HRN · Two-cube collisions on a plane | position MSE | Helped |
| Use k-means groups of 8 particles rather than the tested alternative grouping methods (ablation) | HRN · Grouping-method comparison | prediction accuracy | No clear effect |
| Include training observations with objects at rest (ablation) | HRN · Falling block tower | prediction plausibility | Helped |
Closest papers by meaning
- Visual Grounding of Learned Physical Models (ICML 2020 · similarity 0.84)
- A Compositional Object-Based Approach to Learning Physical Dynamics (ICLR 2017 · similarity 0.83)
- From Neurons to Neutrons: A Case Study in Interpretability (ICML 2024 · similarity 0.81)
- Evaluating alignment between humans and neural network representations in image-based learning tasks (NeurIPS 2024 · similarity 0.80)
- Neural Interpretable PDEs: Harmonizing Fourier Insights with Attention for Scalable and Interpretable Physics Discovery (ICML 2025 · similarity 0.79)
- Homogeneous Dynamics Space for Heterogeneous Humans (CVPR 2025 · similarity 0.78)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.