# Beyond Self-attention: External Attention using Two Linear Layers for Visual Tasks

arXiv: https://arxiv.org/abs/2105.02358 · HTML: https://mlsanity.com/paper/2105.02358/

## Key facts

- Cited by 109 papers in the map (data through 14 Oct 2026).
- Builds on 89 of its 115 references that are in the map.
- First posted on arXiv 5 May 2021.

## Builds on (most cited first, 30 of 89)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [PyTorch: An Imperative Style, High-Performance Deep Learning Library](https://mlsanity.com/paper/1912.01703/) (2019)
- [Microsoft COCO: Common Objects in Context](https://mlsanity.com/paper/1405.0312/) (2014)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks](https://mlsanity.com/paper/1506.01497/) (2015)
- [End-to-End Object Detection with Transformers](https://mlsanity.com/paper/2005.12872/) (2020)
- [Mask R-CNN](https://mlsanity.com/paper/1703.06870/) (2017)
- [Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift](https://mlsanity.com/paper/1502.03167/) (2015)
- [Layer Normalization](https://mlsanity.com/paper/1607.06450/) (2016)
- [PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation](https://mlsanity.com/paper/1612.00593/) (2016)
- [Fully Convolutional Networks for Semantic Segmentation](https://mlsanity.com/paper/1411.4038/) (2014)
- [Squeeze-and-Excitation Networks](https://mlsanity.com/paper/1709.01507/) (2017)
- [Training data-efficient image transformers & distillation through attention](https://mlsanity.com/paper/2012.12877/) (2020)
- [Neural Machine Translation by Jointly Learning to Align and Translate](https://mlsanity.com/paper/1409.0473/) (2014)
- [The Cityscapes Dataset for Semantic Urban Scene Understanding](https://mlsanity.com/paper/1604.01685/) (2016)
- [PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space](https://mlsanity.com/paper/1706.02413/) (2017)
- [Deformable DETR: Deformable Transformers for End-to-End Object Detection](https://mlsanity.com/paper/2010.04159/) (2020)
- [Improved Techniques for Training GANs](https://mlsanity.com/paper/1606.03498/) (2016)
- [XLNet: Generalized Autoregressive Pretraining for Language Understanding](https://mlsanity.com/paper/1906.08237/) (2019)
- [Aggregated Residual Transformations for Deep Neural Networks](https://mlsanity.com/paper/1611.05431/) (2016)
- [Pyramid Scene Parsing Network](https://mlsanity.com/paper/1612.01105/) (2016)
- [Non-local Neural Networks](https://mlsanity.com/paper/1711.07971/) (2017)
- [Improved Training of Wasserstein GANs](https://mlsanity.com/paper/1704.00028/) (2017)
- [Dynamic Graph CNN for Learning on Point Clouds](https://mlsanity.com/paper/1801.07829/) (2018)
- [Unsupervised Representation Learning with Deep Convolutional Generative Adversarial Networks](https://mlsanity.com/paper/1511.06434/) (2015)
- [Pyramid Vision Transformer: A Versatile Backbone for Dense Prediction without Convolutions](https://mlsanity.com/paper/2102.12122/) (2021)
- [BioBERT: a pre-trained biomedical language representation model for biomedical text mining](https://mlsanity.com/paper/1901.08746/) (2019)

## Built on by (most cited first, 30 of 109)

- [SegNeXt: Rethinking Convolutional Attention Design for Semantic Segmentation](https://mlsanity.com/paper/2209.08575/) (2022)
- [Attention Mechanisms in Computer Vision: A Survey](https://mlsanity.com/paper/2111.07624/) (2021)
- [ViTAE: Vision Transformer Advanced by Exploring Intrinsic Inductive Bias](https://mlsanity.com/paper/2106.03348/) (2021)
- [Visual Attention Network](https://mlsanity.com/paper/2202.09741/) (2022)
- [Query2Label: A Simple Transformer Way to Multi-Label Classification](https://mlsanity.com/paper/2107.10834/) (2021)
- [UIU-Net: U-Net in U-Net for Infrared Small Object Detection](https://mlsanity.com/paper/2212.00968/) (2022)
- [CycleMLP: A MLP-like Architecture for Dense Prediction](https://mlsanity.com/paper/2107.10224/) (2021)
- [Vision Permutator: A Permutable MLP-Like Architecture for Visual Recognition](https://mlsanity.com/paper/2106.12368/) (2021)
- [AS-MLP: An Axial Shifted MLP Architecture for Vision](https://mlsanity.com/paper/2107.08391/) (2021)
- [Mixed Transformer U-Net For Medical Image Segmentation](https://mlsanity.com/paper/2111.04734/) (2021)
- [S2-MLP: Spatial-Shift MLP Architecture for Vision](https://mlsanity.com/paper/2106.07477/) (2021)
- [Is Attention Better Than Matrix Decomposition?](https://mlsanity.com/paper/2109.04553/) (2021)
- [OMG-Seg: Is One Model Good Enough For All Segmentation?](https://mlsanity.com/paper/2401.10229/) (2024)
- [Efficient Region-Aware Neural Radiance Fields for High-Fidelity Talking Portrait Synthesis](https://mlsanity.com/paper/2307.09323/) (2023)
- [Delving into Masked Autoencoders for Multi-Label Thorax Disease Classification](https://mlsanity.com/paper/2210.12843/) (2022)
- [Are we ready for a new paradigm shift? A Survey on Visual Deep MLP](https://mlsanity.com/paper/2111.04060/) (2021)
- [On the Connection between Local Attention and Dynamic Depth-wise Convolution](https://mlsanity.com/paper/2106.04263/) (2021)
- [An Image Patch is a Wave: Phase-Aware Vision MLP](https://mlsanity.com/paper/2111.12294/) (2021)
- [MALUNet: A Multi-Attention and Light-weight UNet for Skin Lesion Segmentation](https://mlsanity.com/paper/2211.01784/) (2022)
- [A Survey on Transformer Compression](https://mlsanity.com/paper/2402.05964/) (2024)
- [Sparse MLP for Image Recognition: Is Self-Attention Really Necessary?](https://mlsanity.com/paper/2109.05422/) (2021)
- [SimA: Simple Softmax-free Attention for Vision Transformers](https://mlsanity.com/paper/2206.08898/) (2022)
- [RTFormer: Efficient Design for Real-Time Semantic Segmentation with Transformer](https://mlsanity.com/paper/2210.07124/) (2022)
- [TransLoc3D : Point Cloud based Large-scale Place Recognition using Adaptive Receptive Fields](https://mlsanity.com/paper/2105.11605/) (2021)
- [Eliminating Gradient Conflict in Reference-based Line-Art Colorization](https://mlsanity.com/paper/2207.06095/) (2022)
- [DGR-MIL: Exploring Diverse Global Representation in Multiple Instance Learning for Whole Slide Image Classification](https://mlsanity.com/paper/2407.03575/) (2024)
- [S2-MLPv2: Improved Spatial-Shift MLP Architecture for Vision](https://mlsanity.com/paper/2108.01072/) (2021)
- [Deep Instance Segmentation with Automotive Radar Detection Points](https://mlsanity.com/paper/2110.01775/) (2021)
- [SCTNet: Single-Branch CNN with Transformer Semantic Information for Real-Time Segmentation](https://mlsanity.com/paper/2312.17071/) (2023)
- [AUCSeg: AUC-oriented Pixel-level Long-tail Semantic Segmentation](https://mlsanity.com/paper/2409.20398/) (2024)

## Abstract (arXiv)

Attention mechanisms, especially self-attention, have played an increasingly important role in deep feature representation for visual tasks. Self-attention updates the feature at each position by computing a weighted sum of features using pair-wise affinities across all positions to capture the long-range dependency within a single sample. However, self-attention has quadratic complexity and ignores potential correlation between different samples. This paper proposes a novel attention mechanism which we call external attention, based on two external, small, learnable, shared memories, which can be implemented easily by simply using two cascaded linear layers and two normalization layers; it conveniently replaces self-attention in existing popular architectures. External attention has linear complexity and implicitly considers the correlations between all data samples. We further incorporate the multi-head mechanism into external attention to provide an all-MLP architecture, external attention MLP (EAMLP), for image classification. Extensive experiments on image classification, object detection, semantic segmentation, instance segmentation, image generation, and point cloud analysis reveal that our method provides results comparable or superior to the self-attention mechanism and some of its variants, with much lower computational and memory costs.