Research map / Probabilistic and scientific modeling / Dynamical systems
NeuIDO: Neural Intrinsic Dynamics Operator for Physics-Informed 4D World Models
Jiajing Lin, Xin Zhang, Jianhua Sun
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add few-shot adaptation (100 iterations) to NeuIDO | NeuIDO · NeuMA synthetic scenes | chamfer distance: 1.643 -> 1.216 | Improved |
| add few-shot adaptation to NeuIDO on the Ball scene | NeuIDO · NeuMA synthetic scenes: Ball | chamfer distance: 1.159 -> 1.356 | Hurt |
| increase latent dimension from 32 to 128 (default) in pretraining (ablation) | NeuIDO Stage I · Stage I elasticity MSE | MSE loss: 1.023e3 -> 6.885e2 (dim 32 vs 256) | Improved |
| use latent dimension 256 instead of 128 (ablation) | NeuIDO Stage II · Synthetic scenes alignment | relative CD change: 1.459% average, best or near-best in most scenes | Mixed |
| use latent dimension 32 instead of 128 (ablation) | NeuIDO Stage II · Synthetic scenes alignment | relative CD change: -9.679% | Hurt |
| increase input views from 1 to 4 during Stage II alignment (ablation) | NeuIDO · Synthetic scenes | chamfer distance | Improved |
| increase input views from 4 to 10 during Stage II alignment (ablation) | NeuIDO · Synthetic scenes | chamfer distance | No clear effect |
| use all three views instead of a single view for few-shot adaptation (ablation) | NeuIDO · Real-world scenes: Dog | PSNR: 30.872 -> 34.959 | Improved |
| use all views instead of a single view (ablation) | NeuIDO · Real-world scenes: Pig | PSNR: 33.853 -> 35.295 | Improved |
| add pretraining on constitutive relations (zero-shot) vs random latents | NeuIDO · Unseen PAC-NeRF scenes | PSNR / CD | Improved |
| compare few-shot adaptation across five seeds (ablation) | NeuIDO · Synthetic scenes: Ball | chamfer distance: Mean 1.277, Variance 5.228e-2 | No clear effect |
Closest papers by meaning
- Dyn-O: Building Structured World Models with Object-Centric Representations (NeurIPS 2025 · similarity 0.88)
- Learning 3D Persistent Embodied World Models (NeurIPS 2025 · similarity 0.87)
- World Motion Models: Flexible Sequence Modeling of SE(3) Trajectories (NeurIPS 2026 · similarity 0.87)
- DINO-WM: World Models on Pre-trained Visual Features enable Zero-shot Planning (ICML 2025 · similarity 0.87)
- ODEWorld: A Continuous Predictive Architecture via Physical-Time Flow (NeurIPS 2026 · similarity 0.86)
- NeuMA: Neural Material Adaptor for Visual Grounding of Intrinsic Dynamics (NeurIPS 2024 · similarity 0.86)
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.
