Research map / Architectures and efficiency / RNNs and normalization
WLD-Reg: A Data-Dependent Within-Layer Diversity Regularizer
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 WLD-Reg using the Direct variant | ResNet50 · CIFAR10 | classification error: 7.77 $\pm$ 0.09 vs 8.28 $\pm$ 0.41 | Helped |
| Add WLD-Reg using the Logdet variant | ResNet50 · CIFAR100 | classification error: 32.14 $\pm$ 0.28 -> 31.99 $\pm$ 0.05 | Helped |
| Add WLD-Reg using the Direct variant | ResNet50 · ImageNet-2012 | test error: 23.84 -> 23.24 | Helped |
| Add WLD-Reg using the Det variant | Wide-ResNet50 · ImageNet-2012 | test error: 22.42 -> 21.75 | Helped |
| Add WLD-Reg using the Direct variant | ResNet50 · ImageNet-2012 | generalization gap: 2.87 -> 1.15 | Helped |
| Add WLD-Reg using the Det variant | MLP-Mixer · CIFAR10 | classification error: 23.93 -> 22.66 | Helped |
| Add WLD-Reg using the Logdet variant | gMLP · CIFAR10 | classification error: 22.26 -> 21.56 | Helped |
| Add WLD-Reg using the Logdet variant | ResNet50 · CIFAR10, 40% label noise | classification error | Helped |
| Add WLD-Reg using the Logdet variant | ResNet50 · CIFAR100, 40% label noise | classification error: 48.81 $\pm$ 0.57 -> 46.52 $\pm$ 0.22 | Helped |
| Use weight diversity regularization | ImageNet models · ImageNet-2012 | test error | Mixed |
Closest papers by meaning
- Forward Learning with Top-Down Feedback: Empirical and Analytical Characterization (ICLR 2024 · similarity 0.85)
- Overcoming Rank Collapse in Feedback Alignment (NeurIPS 2026 · similarity 0.84)
- Implicit Regularization in Feedback Alignment Learning Mechanisms for Neural Networks (ICML 2024 · similarity 0.84)
- EUGens: Efficient, Unified and General Dense Layers (NeurIPS 2025 · similarity 0.84)
- Efficient Target Propagation by Deriving Analytical Solution (AAAI 2024 · similarity 0.84)
- The Price of Locality: Why Forward-Forward Underperforms Backpropagation? (NeurIPS 2026 · similarity 0.83)
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.