Deep Convolutional Neural Networks with Zero-Padding: Feature Extraction and Learning

Zhi Han, Baichen Liu, Shao-Bo Lin, Ding-Xuan Zhou
7/30/2023
cs.LGcs.CV

No official code found. Best implementation: liubc17/eDCNN_zero_padding (3 stars, last commit 2023-03-30).

Abstract

This paper studies the performance of deep convolutional neural networks (DCNNs) with zero-padding in feature extraction and learning. After verifying the roles of zero-padding in enabling translation-equivalence, and pooling in its translation-invariance driven nature, we show that with similar number of free parameters, any deep fully connected networks (DFCNs) can be represented by DCNNs with zero-padding. This demonstrates that DCNNs with zero-padding is essentially better than DFCNs in feature extraction. Consequently, we derive universal consistency of DCNNs with zero-padding and show its translation-invariance in the learning process. All our theoretical results are verified by numerical experiments including both toy simulations and real-data running.

View on arXivView PDF

Code Implementations(1)

liubc17/eDCNN_zero_paddingCommunity100%

Deep Convolutional Neural Networks with Zero-Padding: Feature Extraction and Learning

32PythonMar 15, 20233 years ago

Similar Papers

Improving Translation Invariance in Convolutional Neural Networks with Peripheral Prediction Padding

84% match2023

Partial Convolution based Padding

81% match1 repo1.3k2018

zkDL: Efficient Zero-Knowledge Proofs of Deep Learning Training

80% match2023

Feature Learning in Infinite-Width Neural Networks

80% match1 repo612020

Beyond Backpropagation: Monte Carlo Method Can Train Deep Neural Networks

80% match2026

The Master Key Filters Hypothesis: Deep Filters Are General

80% match2024

Cite this paper

@article{han2023deep,
  title  = {Deep Convolutional Neural Networks with Zero-Padding: Feature Extraction and Learning},
  author = {Zhi Han and Baichen Liu and Shao-Bo Lin and Ding-Xuan Zhou},
  year   = {2023},
  eprint = {2307.16203},
  archivePrefix = {arXiv},
  url    = {http://arxiv.org/abs/2307.16203v1}
}

Add a badge to your README

Shows whether official code is linked for this paper.

Code of Paper badge
[![Code of Paper](https://codeofpaper.com/badge/2307.16203.svg)](https://codeofpaper.com/paper/2307.16203)

Discussion