Research map / Architectures and efficiency / RNNs and normalization
CoDiNet: Path Distribution Modeling With Consistency and Diversity for Dynamic Routing
Huanyu Wang, Zequn Qin, Songyuan Li, Xi Li
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 consistency loss to vanilla dynamic routing (ablation) | Vanilla dynamic routing ResNet-110 · CIFAR-10 | accuracy: 93.66 -> 92.88 | Hurt |
| add diversity loss to vanilla dynamic routing (ablation) | Vanilla dynamic routing ResNet-110 · CIFAR-10 | accuracy: 93.66 -> 91.34 | Hurt |
| add both consistency and diversity losses (ablation) | Vanilla dynamic routing ResNet-110 · CIFAR-10 | accuracy: 93.66 -> 94.47 | Improved |
| set diversity margin to 0.5 instead of 0.25 (ablation) | CoDiNet ResNet-110 · CIFAR-10 | accuracy: 93.12 -> 94.47 | Improved |
| increase diversity margin from 0.5 to 0.75 (ablation) | CoDiNet ResNet-110 · CIFAR-10 | accuracy: 94.47 -> 92.75 | Hurt |
| increase diversity margin from 0.5 to 1.00 (ablation) | CoDiNet ResNet-110 · CIFAR-10 | accuracy: 94.47 -> 92.46 | Hurt |
| use CoDiNet instead of ResNet-110 | ResNet-110 · CIFAR-10 | accuracy: 93.60 -> 94.47 | Improved |
| use CoDiNet instead of vanilla dynamic routing | ResNet-110 · CIFAR-100 | accuracy: 70.3 -> 72.9 | Improved |
| use CoDiNet instead of vanilla dynamic routing | ResNet-110 · SVHN | accuracy: 93.15 -> 94.28 | Improved |
| use CoDiNet-50 instead of ResNet-50 | ResNet-50 · ImageNet | top-1 accuracy: 75.36 -> 76.63 | Improved |
| raise cost weight gamma from 0.01 to 0.10 (ablation) | CoDiNet ResNet-110 · CIFAR-10 | accuracy: 94.47 -> 92.45 | Hurt |
| use CoDiNet instead of vanilla dynamic routing | ResNet-32 · CIFAR-10 | accuracy: 91.55 -> 92.48 | Improved |
Later papers that compared with this method
1 later accepted paper reports CoDiNet as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- SP-Net: Slowly Progressing Dynamic Inference Networks (ECCV 2022 · better on 2 of 2 shared results)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MixRoute: Rethinking Single-Distribution Training for Generalizable Neural Routing (NeurIPS 2026 · similarity 0.86)
- Deciding How to Decide: Dynamic Routing in Artificial Neural Networks (ICML 2017 · similarity 0.85)
- Routing Experts: Learning to Route Dynamic Experts in Existing Multi-modal Large Language Models (ICLR 2025 · similarity 0.83)
- Return of ChebNet: Understanding and Improving an Overlooked GNN on Long Range Tasks (NeurIPS 2025 · similarity 0.83)
- Dr.LLM: Dynamic Layer Routing in LLMs (ICLR 2026 · similarity 0.82)
- HyperNet Fields: Efficiently Training Hypernetworks without Ground Truth by Learning Weight Trajectories (CVPR 2025 · 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.
