Efficient Attention Network: Accelerate Attention by Searching Where to Plug

By Zhongzhan Huang, Senwei Liang, Mingfu Liang, Wei He, Haizhao Yang. 2020

12 citations recorded by the source corpus.

Recently, many plug-and-play self-attention modules are proposed to enhance the model generalization by exploiting the internal information of deep convolutional neural networks (CNNs). Previous works lay an emphasis on the design of attention module for specific functionality, e.g., light-weighted or task-oriented attention. However, they ignore the importance of where to plug in the attention module since they connect the modules individually with each block of the entire CNN backbone for granted, leading to incremental computational cost and number of parameters with the growth of network depth. Thus, we propose a framework called Efficient Attention Network (EAN) to improve the efficiency for the existing attention modules. In EAN, we leverage the sharing mechanism (Huang et al. 2020) to share the attention module within the backbone and search where to connect the shared attention module via reinforcement learning. Finally, we obtain the attention network with sparse connections between the backbone and modules, while (1) maintaining accuracy (2) reducing extra parameter increment and (3) accelerating inference. Extensive experiments on widely-used benchmarks and popular attention networks show the effectiveness of EAN. Furthermore, we empirically illustrate that our EAN has the capacity of transferring to other tasks and capturing the informative features. The code is available at https://github.com/gbup-group/EAN-efficient-attention-network

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Wei He
2. Zhongzhan Huang
3. Senwei Liang
4. Ming Liang

In Computer Vision and Pattern Recognition

1. Wei He
2. Zhongzhan Huang
3. Senwei Liang
4. Ming Liang
5. Xiaogang Wang
6. Hongsheng Li
7. Shaoshuai Shi
8. Jiaya Jia

Citations in the corpus

• AlterSGD: Finding Flat Minima for Continual Learning by Alternative Training (2021, background): “Recently, deep neural networks (DNNs) became the dominant paradigm in various real-world applications, for example, image classification [12, 18, 17, 16, 19], semantic segmentation [43, 32, 27], object detection [39, 15, 38, 46, 35].”