Research map / Probabilistic and scientific modeling / Differential equations and PDEs
How to Train Your Neural ODE: the World of Jacobian and Kinetic Regularization
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 |
|---|---|---|---|
| Add kinetic-energy and Jacobian Frobenius-norm regularization | FFJORD · MNIST validation, uniform dequantization | training time: 68.5 -> 24.4 | Helped |
| Add kinetic-energy and Jacobian Frobenius-norm regularization | FFJORD · CIFAR10 validation, uniform dequantization | training time: 91.3 -> 31.8 | Helped |
| Add kinetic-energy and Jacobian Frobenius-norm regularization | FFJORD · MNIST validation, uniform dequantization | bits/dim: 0.97 -> 0.97 | No clear effect |
| Add kinetic-energy and Jacobian Frobenius-norm regularization | FFJORD · CIFAR10 validation, uniform dequantization | bits/dim: 3.36 -> 3.38 | Hurt |
| Add kinetic-energy and Jacobian Frobenius-norm regularization | FFJORD · ImageNet64 | training feasibility | Helped |
| Add Jacobian Frobenius-norm regularization alone (ablation) | FFJORD RNODE training dynamics · MNIST | Jacobian norm | Helped |
| Add kinetic-energy regularization alone (ablation) | FFJORD RNODE training dynamics · MNIST | Jacobian norm | Helped |
| Add kinetic-energy regularization alone (ablation) | FFJORD RNODE training dynamics · MNIST | particle travel distance | Helped |
| Add Jacobian Frobenius-norm regularization alone (ablation) | FFJORD RNODE training dynamics · MNIST | particle travel distance | Helped |
| Remove both regularizers (ablation) | FFJORD RNODE training dynamics · MNIST | training stability | Hurt |
Closest papers by meaning
- How Deep Do We Need: Accelerating Training and Inference of Neural ODEs via Control Perspective (ICML 2024 · similarity 0.91)
- Learning to Solve Generative ODEs Beyond the Linear Span (NeurIPS 2026 · similarity 0.90)
- Global Convergence in Neural ODEs: Impact of Activation Functions (ICLR 2025 · similarity 0.89)
- Simulation-Free Training of Neural ODEs on Paired Data (NeurIPS 2024 · similarity 0.89)
- Semi-Implicit Neural Ordinary Differential Equations (AAAI 2025 · similarity 0.89)
- Improving Neural ODE Training with Temporal Adaptive Batch Normalization (NeurIPS 2024 · similarity 0.88)
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.