Research map / Optimization / Implicit bias and generalization
Deep Frequency Principle Towards Understanding Why Deeper Learning Is Faster
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 a deeper ResNet18 variant | ResNet18 variants · CIFAR10 | training epochs to fixed training accuracy | Helped |
| Use a deeper ResNet18 variant | ResNet18 variants · CIFAR10 | validation accuracy | Helped |
| Increase the depth of the pre-condition component | ResNet18 variants · CIFAR10 effective target function | frequency of RDF concentration during training | Helped |
| Consider a deeper hidden layer in the fixed fully-connected network | Fully connected network · MNIST effective target functions | frequency of RDF peak at the end of training | Helped |
| Use a very deep network without residual connections | Fully connected network · MNIST | training speed | Hurt |
| Add residual connections (ablation) | Very deep fully connected networks · MNIST | training speed across depths | No clear effect |
| Add residual connections (ablation) | Very deep fully connected networks · MNIST effective target function | depth effect on RDF | No clear effect |
Closest papers by meaning
- On the Spectral Bias of Neural Networks (ICML 2019 · similarity 0.86)
- Addressing Spectral Bias of Deep Neural Networks by Multi-Grade Deep Learning (NeurIPS 2024 · similarity 0.84)
- Feature Learning Dynamics in Infinite-Depth Neural Networks (NeurIPS 2026 · similarity 0.83)
- Position: A Theory of Deep Learning Must Include Compositional Sparsity (ICML 2025 · similarity 0.83)
- Analyzing Neural Scaling Laws in Two-Layer Networks with Power-Law Data Spectra (ICLR 2025 · similarity 0.83)
- On the Performance Analysis of Momentum Method: A Frequency Domain Perspective (ICLR 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.