A Novel Convolutional Neural Network Architecture with a Continuous Symmetry
Not on the map. Why?
Abstract (arXiv)
This paper introduces a new Convolutional Neural Network (ConvNet) architecture inspired by a class of partial differential equations (PDEs) called quasi-linear hyperbolic systems. With comparable performance on the image classification task, it allows for the modification of the weights via a continuous group of symmetry. This is a significant shift from traditional models where the architecture and weights are essentially fixed. We wish to promote the (internal) symmetry as a new desirable property for a neural network, and to draw attention to the PDE perspective in analyzing and interpreting ConvNets in the broader Deep Learning community.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 26 of its 36 references that are in the map. First posted on arXiv 3 Aug 2023.
Builds on
26 of 36 references.
References · 26
- 2017Attention Is All You Needcited by 59,932
- 2015Deep Residual Learning for Image Recognitioncited by 51,809
- 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,160
- 2016Densely Connected Convolutional Networkscited by 6,691
- 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,827
- 2018Neural Ordinary Differential Equationscited by 3,733
- 2022A ConvNet for the 2020scited by 3,545
- 2020Fourier Neural Operator for Parametric Partial Differential Equationscited by 2,806
- 2016Xception: Deep Learning with Depthwise Separable Convolutionscited by 2,320
- 2020GLU Variants Improve Transformercited by 1,724
- 2016Group Equivariant Convolutional Networkscited by 1,154
- 2021Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gaugescited by 1,131
14 more
- 2016Language Modeling with Gated Convolutional Networkscited by 688
- 2018Efficient Neural Network Robustness Certification with General Activation Functionscited by 439
- 2021ResNet strikes back: An improved training procedure in timmcited by 365
- 2017PDE-Net: Learning PDEs from Datacited by 304
- 2018Deep Neural Networks Motivated by Partial Differential Equationscited by 187
- 2022Scaling Up Your Kernels to 31x31: Revisiting Large Kernel Design in CNNscited by 155
- 2021Diverse Branch Block: Building a Convolution as an Inception-like Unitcited by 87
- 2020Algorithms for Solving High Dimensional PDEs: From Nonlinear Monte Carlo to Machine Learningcited by 77
- 2020Deep Polynomial Neural Networkscited by 63
- 2020Why Attentions May Not Be Interpretable?cited by 41
- 2020Pi-nets: Deep Polynomial Neural Networkscited by 39
- 2022Symmetries, flat minima, and the conserved quantities of gradient flowcited by 35
- 2021Continuous vs. Discrete Optimization of Deep Neural Networkscited by 34
- 2022Do Residual Neural Networks discretize Neural Ordinary Differential Equations?cited by 27
Built on by
Citing papers per year: 2023: 1; 2024: 0; 2025: 1.
2 papers.