Research map / Probabilistic and scientific modeling / Differential equations and PDEs
Neural Delay Differential Equations
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 NDDE instead of NODE | continuous-depth image classifier · CIFAR10 | test accuracy: 53.92%± 0.67 -> 55.69%± 0.39 | Helped |
| Use NDDE instead of NODE | continuous-depth image classifier · MNIST | test accuracy: 96.21%± 0.66 -> 96.22%± 0.55 | No clear effect |
| Use NDDE instead of NODE | continuous-depth image classifier · SVHN | test accuracy: 80.66%± 0.56 -> 81.49%± 0.09 | Helped |
| Add an ODE-modeled initial function to NDDE | NDDE image classifier · CIFAR10 | test accuracy: 55.69%± 0.39 -> ${\bf 55.89\%\pm 0.71}$ | Helped |
| Add an ODE-modeled initial function to NDDE | NDDE image classifier · MNIST | test accuracy: 96.22%± 0.55 -> ${\bf 97.26\%\pm 0.22}$ | Helped |
| Add an ODE-modeled initial function to NDDE | NDDE image classifier · SVHN | test accuracy: 81.49%± 0.09 -> ${\bf 82.60\%\pm 0.22}$ | Helped |
| Add delay to NODE | NODE time-series model · synthetic delayed time series, population dynamics, and Mackey-Glass | training/test loss | Helped |
| Add delayed state information | image classifier · CIFAR10, MNIST, and SVHN | test accuracy | Helped |
| Augment image space by 1 channel (ablation) | NODE · CIFAR10 | test accuracy: 53.92%± 0.67 -> 56.14%± 0.48 | Helped |
| Augment image space by 1 channel (ablation) | NDDE · MNIST | test accuracy: 96.22%± 0.55 -> 97.83%± 0.07 | Helped |
| Augment image space by 2 channels (ablation) | NODE+NDDE · SVHN | test accuracy: ${\bf 82.60\%\pm 0.22}$ -> ${\bf 83.73\%\pm 0.72}$ | Helped |
| Augment image space by 4 channels (ablation) | NDDE · MNIST | test accuracy: 96.22%± 0.55 -> 98.31%± 0.03 | Helped |
Closest papers by meaning
- Neural Piecewise-Constant Delay Differential Equations (AAAI 2022 · similarity 0.92)
- Neural Differential Equations for Learning to Program Neural Nets Through Continuous Learning Rules (NeurIPS 2022 · similarity 0.90)
- On the Generalization and Approximation Capacities of Neural Controlled Differential Equations (ICLR 2024 · similarity 0.87)
- Symbolic Neural Ordinary Differential Equations (AAAI 2025 · similarity 0.87)
- 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)
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.