A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf
.
Attention, please! A survey of Neural Attention Models in Deep Learning
[article]
2021
arXiv
pre-print
In humans, Attention is a core property of all perceptual and cognitive operations. Given our limited ability to process competing sources, attention mechanisms select, modulate, and focus on the information most relevant to behavior. For decades, concepts and functions of attention have been studied in philosophy, psychology, neuroscience, and computing. For the last six years, this property has been widely explored in deep neural networks. Currently, the state-of-the-art in Deep Learning is
arXiv:2103.16775v1
fatcat:lwkw42lrircorkymykpgdmlbwq