Research map / Probabilistic and scientific modeling / Dynamical systems
Dynamical System Inspired Adaptive Time Stepping Controller for Residual Network Families
Yibo Yang, Jianlong Wu, Hongyang Li, Xia Li, Tiancheng Shen, Zhouchen Lin
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 the adaptive time stepping controller | ResNet-18 · ImageNet val | top-1 error (%): 29.41 -> 28.80 | Improved |
| Add the adaptive time stepping controller | ResNet-34 · ImageNet val | top-1 error (%): 26.03 -> 25.39 | Improved |
| Add the adaptive time stepping controller | ResNet-50 · ImageNet val | top-1 error (%): 24.42 -> 23.63 | Improved |
| Add the adaptive time stepping controller | ResNet-101 · ImageNet val | top-1 error (%): 22.94 -> 22.24 | Improved |
| Add the adaptive time stepping controller | ResNeXt-50 · ImageNet val | top-1 error (%): 22.84 -> 22.23 | Improved |
| Add the adaptive time stepping controller | ResNeXt-101 · ImageNet val | top-1 error (%): 21.88 -> 21.17 | Improved |
| Add the adaptive time stepping controller | SENet-50 · ImageNet val | top-1 error (%): 23.27 -> 22.75 | Improved |
| Add the adaptive time stepping controller | SENet-101 · ImageNet val | top-1 error (%): 22.37 -> 21.82 | Improved |
| Add the adaptive time stepping controller | DenseNet · CIFAR-10 | error rate: 5.24 -> 5.10 | Improved |
| Add the adaptive time stepping controller | DenseNet · CIFAR-100 | error rate: 24.42 -> 23.35 | Improved |
| Add the adaptive time stepping controller | CliqueNet · CIFAR-10 | error rate: 5.03 -> 4.79 | Improved |
| Add the adaptive time stepping controller | CliqueNet · CIFAR-100 | error rate: 22.80 -> 22.04 | Improved |
Closest papers by meaning
- Towards Adaptive Residual Network Training: A Neural-ODE Perspective (ICML 2020 · similarity 0.84)
- Reinforcing Neural Network Stability with Attractor Dynamics (AAAI 2020 · similarity 0.84)
- CSG-ODE: ControlSynth Graph ODE For Modeling Complex Evolution of Dynamic Graphs (ICML 2025 · similarity 0.83)
- Transformative or Conservative? Conservation laws for ResNets and Transformers (ICML 2025 · similarity 0.82)
- Leveraging Low-Rank and Sparse Recurrent Connectivity for Robust Closed-Loop Control (ICLR 2024 · similarity 0.82)
- One-Time Soft Alignment Enables Resilient Learning without Weight Transport (NeurIPS 2026 · similarity 0.82)
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.
