# Co-Attentive Equivariant Neural Networks: Focusing Equivariance On Transformations Co-Occurring In Data

@article{Romero2020CoAttentiveEN, title={Co-Attentive Equivariant Neural Networks: Focusing Equivariance On Transformations Co-Occurring In Data}, author={David W. Romero and Mark Hoogendoorn}, journal={ArXiv}, year={2020}, volume={abs/1911.07849} }

Equivariance is a nice property to have as it produces much more parameter efficient neural architectures and preserves the structure of the input through the feature mapping. Even though some combinations of transformations might never appear (e.g. an upright face with a horizontal nose), current equivariant architectures consider the set of all possible transformations in a transformation group when learning feature representations. Contrarily, the human visual system is able to attend to the… Expand

#### 14 Citations

Group Equivariant Subsampling

- Computer Science, Mathematics
- ArXiv
- 2021

Group equivariant autoencoders (GAEs) are used in models that learn object-centric representations on multiobject datasets, and show improved data efficiency and decomposition compared to non-equivariant baselines. Expand

Group Equivariant Neural Architecture Search via Group Decomposition and Reinforcement Learning

- Computer Science
- ArXiv
- 2021

It is shown that AENs find the right balance between group equivariance and number of parameters, thereby consistently having good task performance, yielding what the authors call autoequivariant networks (AENs). Expand

Equivariant Wavelets: Fast Rotation and Translation Invariant Wavelet Scattering Transforms

- Computer Science, Physics
- ArXiv
- 2021

This work introduces a fast-to-compute, translationally invariant and rotationally equivariant wavelet scattering network (EqWS) and filter bank of wavelets (triglets) and demonstrates the interpretability and quantify the invariance/equivariance of the coefficients. Expand

Exploiting Learned Symmetries in Group Equivariant Convolutions

- Computer Science
- ICIP 2021
- 2021

It is shown that GConvs can be efficiently decomposed into depthwise separable convolutions while preserving equivariance properties and demonstrated improved performance and data efficiency on two datasets. Expand

A Dynamic Group Equivariant Convolutional Networks for Medical Image Analysis

- Computer Science
- 2020 IEEE International Conference on Bioinformatics and Biomedicine (BIBM)
- 2020

This paper proposes a generalization of the dynamic Convolutional method, named as dynamic group equivariant convolution, to strengthen the relationships and increase model capability by aggregating multiple group convolutional kernels via attention, and demonstrates that breast tumor classification is substantial improvements when compared to a recent baseline architecture. Expand

Wavelet Networks: Scale Equivariant Learning From Raw Waveforms

- Computer Science, Mathematics
- ArXiv
- 2020

This work utilizes the concept of scale and translation equivariance to tackle the problem of learning on time-series from raw waveforms, and obtains representations that largely resemble those of the wavelet transform at the first layer, but that evolve into much more descriptive ones as a function of depth. Expand

Boosting Deep Neural Networks with Geometrical Prior Knowledge: A Survey

- Computer Science
- ArXiv
- 2020

This survey tries to give a concise overview about different approaches to incorporate geometrical prior knowledge into DNNs, and tries to connect those methods to the field of 3D object detection for autonomous driving, where they expect promising results applying those methods. Expand

Attentive Group Equivariant Convolutional Networks

- Computer Science, Mathematics
- ICML
- 2020

Attentive group equivariant convolutions are presented, a generalization of the group convolution, in which attention is applied during the course of convolution to accentuate meaningful symmetry combinations and suppress non-plausible, misleading ones. Expand

Autoequivariant Network Search via Group Decomposition

- Computer Science
- 2021

Recent works show that group equivariance as an inductive bias improves neural network performance for both classification and generation. However, designing group-equivariant neural networks is… Expand

CNN Architectures for Geometric Transformation-Invariant Feature Representation in Computer Vision: A Review

- Computer Science
- SN Comput. Sci.
- 2021

#### References

SHOWING 1-10 OF 42 REFERENCES

Harmonic Networks: Deep Translation and Rotation Equivariance

- Computer Science, Mathematics
- 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
- 2017

H-Nets are presented, a CNN exhibiting equivariance to patch-wise translation and 360-rotation, and it is demonstrated that their layers are general enough to be used in conjunction with the latest architectures and techniques, such as deep supervision and batch normalization. Expand

Learning Steerable Filters for Rotation Equivariant CNNs

- Computer Science, Mathematics
- 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition
- 2018

Steerable Filter CNNs (SFCNNs) are developed which achieve joint equivariance under translations and rotations by design and generalize He's weight initialization scheme to filters which are defined as a linear combination of a system of atomic filters. Expand

Deep Rotation Equivariant Network

- Computer Science
- Neurocomputing
- 2018

This work proposes Deep Rotation Equivariant Network consisting of cycle layers, isotonic layers and decycle layers, and evaluates DRENs on Rotated MNIST and CIFAR-10 datasets and demonstrates that it can improve the performance of state-of-the-art architectures. Expand

Extracting Invariant Features From Images Using An Equivariant Autoencoder

- Mathematics, Computer Science
- ACML
- 2018

This work applies group convolutions to build an Equivariant Autoencoder with embeddings that change predictably under the specified set of transformations, and introduces two approaches to extracting invariant features from theseembeddings—Gram Pooling and Equivariants Attention. Expand

Rotation Equivariant Vector Field Networks

- Computer Science, Mathematics
- 2017 IEEE International Conference on Computer Vision (ICCV)
- 2017

The Rotation Equivariant Vector Field Networks (RotEqNet), a Convolutional Neural Network architecture encoding rotation equivariance, invariance and covariance, is proposed and a modified convolution operator relying on this representation to obtain deep architectures is developed. Expand

Gauge Equivariant Convolutional Networks and the Icosahedral CNN

- Computer Science, Mathematics
- ICML
- 2019

Gauge equivariant convolution using a single conv2d call is demonstrated, making it a highly scalable and practical alternative to Spherical CNNs and demonstrating substantial improvements over previous methods on the task of segmenting omnidirectional images and global climate patterns. Expand

Deep Symmetry Networks

- Computer Science, Mathematics
- NIPS
- 2014

Deep symmetry networks (symnets), a generalization of convnets that forms feature maps over arbitrary symmetry groups that uses kernel-based interpolation to tractably tie parameters and pool over symmetry spaces of any dimension are introduced. Expand

Scale equivariance in CNNs with vector fields

- Computer Science
- ArXiv
- 2018

This work studies the effect of injecting local scale equivariance into Convolutional Neural Networks and shows that this improves the performance of the model by over 20% in the scale Equivariant task of regressing the scaling factor applied to randomly scaled MNIST digits. Expand

CubeNet: Equivariance to 3D Rotation and Translation

- Computer Science, Mathematics
- ECCV
- 2018

A Group Convolutional Neural Network with linear equivariance to translations and right angle rotations in three dimensions is introduced, and is believed to be the first 3D rotation equivariant CNN for voxel representations. Expand

Exploiting Cyclic Symmetry in Convolutional Neural Networks

- Computer Science
- ICML
- 2016

This work introduces four operations which can be inserted into neural network models as layers, andWhich can be combined to make these models partially equivariant to rotations, and which enable parameter sharing across different orientations. Expand