Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Augmented Neural ODEs
Emilien Dupont, Arnaud Doucet, Yee Whye Teh
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 augmented dimensions to the Neural ODE | Neural ODE · MNIST test | accuracy: 96.4% ± 0.5 -> 98.2% ± 0.1 | Improved |
| Add augmented dimensions to the Neural ODE | Neural ODE · CIFAR10 test | accuracy: 53.7% ± 0.2 -> 60.6% ± 0.4 | Improved |
| Add augmented dimensions to the Neural ODE | Neural ODE · SVHN test | accuracy: 81.0% ± 0.6 -> 83.5% ± 0.5 | Improved |
| Add 10 augmented channels | Neural ODE · MNIST | training iterations to reach the same loss: roughly 10 times fewer iterations | Improved |
| Add augmented dimensions | Neural ODE · CIFAR10 | training iterations: roughly 5 times faster | Improved |
| Use augmentation rather than an unaugmented ODE | Neural ODE · CIFAR10 | number of function evaluations at loss 0.8: approximately 100 -> 50 | Improved |
| Combine augmentation with weight decay (ablation) | Augmented Neural ODE · image datasets | number of function evaluations | Improved |
| Combine augmentation with weight decay (ablation) | Augmented Neural ODE · image datasets | loss: slightly higher loss | Hurt |
| Use excessively large augmentation, such as adding 100 channels (ablation) | Augmented Neural ODE · MNIST | loss | Hurt |
| Use excessively large augmentation, such as adding 100 channels (ablation) | Augmented Neural ODE · MNIST | number of function evaluations | Hurt |
| Use softplus instead of ReLU (ablation) | Neural ODE and Augmented Neural ODE · experiments | learning speed | Hurt |
Closest papers by meaning
- On Second Order Behaviour in Augmented Neural ODEs (NeurIPS 2020 · similarity 0.89)
- Semi-Implicit Neural Ordinary Differential Equations (AAAI 2025 · similarity 0.88)
- How Deep Do We Need: Accelerating Training and Inference of Neural ODEs via Control Perspective (ICML 2024 · similarity 0.87)
- Neural ODE Transformers: Analyzing Internal Dynamics and Adaptive Fine-tuning (ICLR 2025 · similarity 0.86)
- Global Convergence in Neural ODEs: Impact of Activation Functions (ICLR 2025 · similarity 0.86)
- Neural Conjugate Flows: A Physics-Informed Architecture with Flow Structure (AAAI 2025 · similarity 0.85)
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.
