Research map / Learning theory and algorithms / Sample complexity
Deep Network Approximation: Achieving Arbitrary Accuracy with Fixed Number of Neurons
Shijun Zhang, Zuowei Shen, Haizhao Yang
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 |
|---|---|---|---|
| replace ReLU by EUAF activation in a 4x80 fully connected network | FCNN1 · Function approximation (oscillatory 2D target) | test MSE: 3.53e-5 -> 7.56e-6 | Improved |
| replace ReLU by EUAF in a 4x80 fully connected network | FCNN1 · Function approximation (oscillatory 2D target) | test MAE: 4.57e-3 -> 2.13e-3 | Improved |
| replace ReLU by EUAF in a 4x80 fully connected network | FCNN1 · Function approximation (oscillatory 2D target) | test MAX loss: 3.69e-2 -> 1.48e-2 | Improved |
| replace one channel or neuron per hidden layer of ReLU by EUAF | CNN1 · Fashion-MNIST test | average accuracy in last 100 epochs: 0.932698 -> 0.933508 | Improved |
| replace one channel or neuron per hidden layer of ReLU by EUAF | CNN1 · Fashion-MNIST test | largest test accuracy: 0.933066 -> 0.933922 | Improved |
| replace one channel or neuron per hidden layer of ReLU by EUAF | CNN1 · Fashion-MNIST test | average of largest 100 accuracies: 0.932852 -> 0.933685 | Improved |
Closest papers by meaning
- Deep Network Approximation: Beyond ReLU to Diverse Activation Functions (JMLR 2024 · similarity 0.90)
- Why Deep Neural Networks for Function Approximation? (ICLR 2017 · similarity 0.89)
- Approximation and non-parametric estimation of functions over high-dimensional spheres via deep ReLU networks (ICLR 2023 · similarity 0.88)
- Minimum width for universal approximation using ReLU networks on compact domain (ICLR 2024 · similarity 0.86)
- Characterizing ResNet's Universal Approximation Capability (ICML 2024 · similarity 0.86)
- ReLU Network with Width $d+\mathcal{O}(1)$ Can Achieve Optimal Approximation Rate (ICML 2024 · similarity 0.86)
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.
