Journal article
Tonio Weidler, Julian Lehnen, Quinton Denman, Dávid Sebok, Gerhard Weiss, K. Driessens, M. Senden
arXiv.org, 2021
arXiv.org, 2021
Semantic Scholar
ArXiv
DBLP
Cite
Cite
APA
Click to copy
Weidler, T., Lehnen, J., Denman, Q., Sebok, D., Weiss, G., Driessens, K., & Senden, M. (2021). Biologically Inspired Semantic Lateral Connectivity for Convolutional Neural Networks. ArXiv.org.
Chicago/Turabian
Click to copy
Weidler, Tonio, Julian Lehnen, Quinton Denman, Dávid Sebok, Gerhard Weiss, K. Driessens, and M. Senden. “Biologically Inspired Semantic Lateral Connectivity for Convolutional Neural Networks.” arXiv.org (2021).
MLA
Click to copy
Weidler, Tonio, et al. “Biologically Inspired Semantic Lateral Connectivity for Convolutional Neural Networks.” ArXiv.org, 2021.
BibTeX Click to copy
@article{tonio2021a,
title = {Biologically Inspired Semantic Lateral Connectivity for Convolutional Neural Networks},
year = {2021},
journal = {arXiv.org},
author = {Weidler, Tonio and Lehnen, Julian and Denman, Quinton and Sebok, Dávid and Weiss, Gerhard and Driessens, K. and Senden, M.}
}
Abstract
Lateral connections play an important role for sensory processing in visual cortex by supporting discriminable neuronal responses even to highly similar features. In the present work, we show that establishing a biologically inspired Mexican hat lateral connectivity profile along the filter domain can significantly improve the classification accuracy of a variety of lightweight convolutional neural networks without the addition of trainable network parameters. Moreover, we demonstrate that it is possible to analytically determine the stationary distribution of modulated filter activations and thereby avoid using recurrence for modeling temporal dynamics. We furthermore reveal that the Mexican hat connectivity profile has the effect of ordering filters in a sequence resembling the topographic organization of feature selectivity in early visual cortex. In an ordered filter sequence, this profile then sharpens the filters’ tuning curves.