Research map / Architectures and efficiency / CNN design
Inductive Bias of Deep Convolutional Networks through Pooling Geometry
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 square rather than mirror pooling | Deep convolutional arithmetic circuit · Synthetic closedness classification | accuracy | Helped |
| use mirror rather than square pooling | Deep convolutional arithmetic circuit · Synthetic symmetry classification | accuracy | Helped |
| use a shallow rather than deep architecture | Convolutional arithmetic circuit · Synthetic closedness classification | accuracy | Hurt |
| use a shallow rather than deep architecture | Convolutional arithmetic circuit · Synthetic symmetry classification | accuracy | Hurt |
| use a shallow rather than deep architecture | Convolutional rectifier network with average pooling · Synthetic closedness classification | accuracy | Hurt |
| use a shallow rather than deep architecture | Convolutional rectifier network with average pooling · Synthetic symmetry classification | accuracy | Hurt |
| use max pooling in the shallow rectifier network | Shallow convolutional rectifier network · Synthetic closedness and symmetry classification | accuracy | Hurt |
Closest papers by meaning
- Theoretical Analysis of the Inductive Biases in Deep Convolutional Networks (NeurIPS 2023 · similarity 0.89)
- Approximation and Learning with Deep Convolutional Models: a Kernel Perspective (ICLR 2022 · similarity 0.88)
- Structured Convolutions for Efficient Neural Network Design (NeurIPS 2020 · similarity 0.88)
- TI-Pooling: Transformation-Invariant Pooling for Feature Learning in Convolutional Neural Networks (CVPR 2016 · similarity 0.87)
- Maximum Class Separation as Inductive Bias in One Matrix (NeurIPS 2022 · similarity 0.87)
- Unveiling the Unseen: Identifiable Clusters in Trained Depthwise Convolutional Kernels (ICLR 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.