Research map / Architectures and efficiency / RNNs and normalization
On the Expressive Power of Deep Polynomial Neural Networks
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 |
|---|---|---|---|
| Increase activation degree | Polynomial networks with fixed architecture · Examples of architectures in the dimensions table | functional variety dimension: 5 -> 6; 6 -> 6; 6 -> 6; 6 -> 6 | Mixed |
| Use intermediate widths that increase and then decrease | Polynomial networks seeking a filling functional variety · Computed minimal filling architectures | filling functional variety | Helped |
| Constrain an intermediate width to 2d0 - 2 | Deep polynomial networks · Asymptotic depth, r >= 2 and d0 >= 2 | filling functional variety | Hurt |
Closest papers by meaning
- Identifiability of Deep Polynomial Neural Networks (NeurIPS 2025 · similarity 0.91)
- Polynomial Width is Sufficient for Set Representation with High-dimensional Features (ICLR 2024 · similarity 0.88)
- Position: Algebra Unveils Deep Learning - An Invitation to Neuroalgebraic Geometry (ICML 2025 · similarity 0.86)
- How Universal Polynomial Bases Enhance Spectral Graph Neural Networks: Heterophily, Over-smoothing, and Over-squashing (ICML 2024 · similarity 0.86)
- On the Expressiveness of Rational ReLU Neural Networks With Bounded Depth (ICLR 2025 · similarity 0.86)
- Graph Neural Networks and Arithmetic Circuits (NeurIPS 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.