Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Heavy Ball Neural Ordinary Differential Equations
Hedi Xia, Vai Suliafu, Hangjie Ji, Tan Minh Nguyen, Andrea Bertozzi, Stanley Osher, Bao Wang
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 |
|---|---|---|---|
| Replace first-order NODE with heavy-ball (momentum) dynamics HBNODE/GHBNODE | NODE · Point cloud separation | training loss and NFEs | Improved |
| Use HBNODE instead of ANODE/SONODE augmentation | ANODE, SONODE · Point cloud separation | training loss reaching zero | Improved |
| Plain NODE on point clouds | NODE · Point cloud separation | separation success | Hurt |
| Use HBNODE/GHBNODE on MNIST and CIFAR10 | NODE, ANODE, SONODE · MNIST, CIFAR10 | forward and backward NFEs | Improved |
| Input augmentation (ANODE, SONODE) vs plain NODE | NODE · MNIST, CIFAR10 | forward and backward NFEs | Mixed |
| Use HBNODE/GHBNODE | NODE, ANODE, SONODE · MNIST, CIFAR10 | test accuracy | Improved |
| Decrease ODE solver tolerance | HBNODE vs NODE/ANODE/SONODE · CIFAR10 | NFEs | Improved |
| Use HBNODE-RNN/GHBNODE-RNN instead of ODE-RNN/ANODE-RNN | ODE-RNN, ANODE-RNN · Plane vibration | train and test loss | Improved |
| Use HBNODE-RNN/GHBNODE-RNN instead of SONODE-RNN | SONODE-RNN · Plane vibration | prediction quality and backward NFEs | Improved |
| Use HBNODE-RNN/GHBNODE-RNN instead of ODE-RNN/ANODE-RNN | ODE-RNN, ANODE-RNN · Walker2D | NFEs and train/test loss | Improved |
| SONODE-RNN on Walker2D | SONODE-RNN · Walker2D | trainability | Hurt |
Closest papers by meaning
- Improving Neural Ordinary Differential Equations with Nesterov's Accelerated Gradient Method (NeurIPS 2022 · similarity 0.89)
- Efficient Certified Training and Robustness Verification of Neural ODEs (ICLR 2023 · similarity 0.89)
- Opening the Blackbox: Accelerating Neural Differential Equations by Regularizing Internal Solver Heuristics (ICML 2021 · similarity 0.89)
- ControlSynth Neural ODEs: Modeling Dynamical Systems with Guaranteed Convergence (NeurIPS 2024 · similarity 0.87)
- How Deep Do We Need: Accelerating Training and Inference of Neural ODEs via Control Perspective (ICML 2024 · similarity 0.87)
- Semi-Implicit Neural Ordinary Differential Equations (AAAI 2025 · similarity 0.87)
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.
