Research map / Architectures and efficiency / RNNs and normalization
ANODEV2: A Coupled Neural ODE Framework
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 evolving weights via coupled ODE (config 1) | AlexNet baseline · CIFAR-10 | test accuracy (avg): 87.03% -> 88.78% | Helped |
| add evolving weights via coupled ODE (config 1) | ResNet-4 baseline · CIFAR-10 | test accuracy (avg): 76.95% -> 78.11% | Helped |
| add evolving weights via coupled ODE (config 1) | ResNet-10 baseline · CIFAR-10 | test accuracy (avg): 88.10% -> 89.04% | Helped |
| evolve weights in time instead of static weights (same FLOPs and size, 2 activation steps) (ablation) | ANODE · CIFAR-10 AlexNet | test accuracy (avg): 88.02% -> 88.26% | Helped |
| evolve weights in time instead of static weights (same FLOPs and size, 2 activation steps) (ablation) | ANODE · CIFAR-10 ResNet-4 | test accuracy (avg): 77.30% -> 77.73% | Helped |
| evolve weights in time instead of static weights (same FLOPs and size, 2 activation steps) (ablation) | ANODE · CIFAR-10 ResNet-10 | test accuracy (avg): 88.60% -> 88.93% | Helped |
| use adjoint Neural ODE instead of discretize-then-optimize ANODE | ANODE · CIFAR-10 ResNet-4 | test accuracy (avg): 77.30% -> 47.37% | Hurt |
| config 2 (parameters applied after fixed steps) instead of config 1 (ablation) | ANODEV2 · CIFAR-10 AlexNet | test accuracy (avg): 88.78% -> 88.26% | Mixed |
Closest papers by meaning
- Do Residual Neural Networks discretize Neural Ordinary Differential Equations? (NeurIPS 2022 · similarity 0.91)
- Implicit regularization of deep residual networks towards neural ODEs (ICLR 2024 · similarity 0.89)
- Simulation-Free Training of Neural ODEs on Paired Data (NeurIPS 2024 · similarity 0.89)
- How Deep Do We Need: Accelerating Training and Inference of Neural ODEs via Control Perspective (ICML 2024 · similarity 0.88)
- Improving Neural ODE Training with Temporal Adaptive Batch Normalization (NeurIPS 2024 · similarity 0.88)
- Global Convergence in Neural ODEs: Impact of Activation Functions (ICLR 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.