arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2102.12895cs.LG
收藏

用残差卷积演化注意力

Evolving Attention with Residual Convolutions

Peking University(北京大学);Microsoft Research(微软研究院);Institute of Information Engineering, Chinese Academy of Sciences(中国科学院信息工程研究所);ETH Zurich(苏黎世联邦理工学院);Tsinghua University(清华大学)

Yujing Wang, Yaming Yang, Jiangang Bai, Mingliang Zhang, Jing Bai, Jing Yu, Ce Zhang, Gao Huang, Yunhai Tong

摘要:

Transformer是自然语言处理中一种普遍存在的模型,在计算机视觉中引起了广泛的关注。注意力图对于Transformer模型编码输入标记之间的依赖关系是必不可少的。然而,它们在每一层都是独立学习的,有时无法捕捉精确的模式。在本文中,我们提出了一种新的和通用的机制的基础上不断发展的注意力,以提高Transformers的性能。一方面,不同层次的注意力图共享公共知识,因此前几层的注意力图可以通过剩余连接来指导后几层的注意力。另一方面,低级别和高级别注意力在抽象层次上是不同的,因此我们采用卷积层来模拟注意力地图的进化过程。所提出的进化注意力机制实现了显着的性能改善,在各种国家的最先进的模型,多个任务,包括图像分类,自然语言理解和机器翻译。

英文摘要:

Transformer is a ubiquitous model for natural language processing and has attracted wide attentions in computer vision. The attention maps are indispensable for a transformer model to encode the dependencies among input tokens. However, they are learned independently in each layer and sometimes fail to capture precise patterns. In this paper, we propose a novel and generic mechanism based on evolving attention to improve the performance of transformers. On one hand, the attention maps in different layers share common knowledge, thus the ones in preceding layers can instruct the attention in succeeding layers through residual connections. On the other hand, low-level and high-level attentions vary in the level of abstraction, so we adopt convolutional layers to model the evolutionary process of attention maps. The proposed evolving attention mechanism achieves significant performance improvement over various state-of-the-art models for multiple tasks, including image classification, natural language understanding and machine translation.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑