Evolution: A Unified Formula for Feature Operators from a High-level Perspective
Not on the map. Why?
Abstract (arXiv)
Traditionally, different types of feature operators (e.g., convolution, self-attention and involution) utilize different approaches to extract and aggregate the features. Resemblance can be hardly discovered from their mathematical formulas. However, these three operators all serve the same paramount purpose and bear no difference in essence. Hence we probe into the essence of various feature operators from a high-level perspective, transformed their components equivalently, and explored their mathematical expressions within higher dimensions. We raise one clear and concrete unified formula for different feature operators termed as Evolution. Evolution utilizes the Evolution Function to generate the Evolution Kernel, which extracts and aggregates the features in certain positions of the input feature map. We mathematically deduce the equivalent transformation from the traditional formulas of these feature operators to Evolution and prove the unification. In addition, we discuss the forms of Evolution Functions and the properties of generated Evolution Kernels, intending to give inspirations to the further research and innovations of powerful feature operators.
Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Builds on 26 of its 29 references that are in the map. First posted on arXiv 23 May 2023.
Builds on
26 of 29 references.
References · 26
- 2017Attention Is All You Needcited by 59,932
- 2015Deep Residual Learning for Image Recognitioncited by 51,809
- 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
- 2014Very Deep Convolutional Networks for Large-Scale Image Recognitioncited by 15,267
- 2021Swin Transformer: Hierarchical Vision Transformer using Shifted Windowscited by 12,082
- 2016Densely Connected Convolutional Networkscited by 6,691
- 2018MobileNetV2: Inverted Residuals and Linear Bottleneckscited by 4,773
- 2017MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applicationscited by 3,827
- 2016Xception: Deep Learning with Depthwise Separable Convolutionscited by 2,320
- 2017Deformable Convolutional Networkscited by 1,481
- 2021MLP-Mixer: An all-MLP Architecture for Visioncited by 1,441
- 2015Multi-Scale Context Aggregation by Dilated Convolutionscited by 1,128
14 more
- 2021CoAtNet: Marrying Convolution and Attention for All Data Sizescited by 566
- 2019Stand-Alone Self-Attention in Vision Modelscited by 542
- 2019Dynamic Convolution: Attention over Convolution Kernelscited by 355
- 2019Axial Attention in Multidimensional Transformerscited by 350
- 2021ResMLP: Feedforward networks for image classification with data-efficient trainingcited by 338
- 2019On the Relationship between Self-Attention and Convolutional Layerscited by 299
- 2019CondConv: Conditionally Parameterized Convolutions for Efficient Inferencecited by 273
- 2019Local Relation Networks for Image Recognitioncited by 210
- 2021Involution: Inverting the Inherence of Convolution for Visual Recognitioncited by 99
- 2021Do You Even Need Attention? A Stack of Feed-Forward Layers Does Surprisingly Well on ImageNetcited by 61
- 2022Omni-Dimensional Dynamic Convolutioncited by 55
- 2020WeightNet: Revisiting the Design Space of Weight Networkscited by 50
- 2020DyNet: Dynamic Convolution for Accelerating Convolutional Neural Networkscited by 39
- 2021X-volution: On the unification of convolution and self-attentioncited by 9
Built on by
Citing papers per year: 2025: 2.
2 papers.