Cargando…

The geometry of robustness in spiking neural networks

Neural systems are remarkably robust against various perturbations, a phenomenon that still requires a clear explanation. Here, we graphically illustrate how neural networks can become robust. We study spiking networks that generate low-dimensional representations, and we show that the neurons’ subt...

Descripción completa

Detalles Bibliográficos
Autores principales: Calaim, Nuno, Dehmelt, Florian A, Gonçalves, Pedro J, Machens, Christian K
Formato: Online Artículo Texto
Lenguaje:English
Publicado: eLife Sciences Publications, Ltd 2022
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9307274/
https://www.ncbi.nlm.nih.gov/pubmed/35635432
http://dx.doi.org/10.7554/eLife.73276
_version_ 1784752721441587200
author Calaim, Nuno
Dehmelt, Florian A
Gonçalves, Pedro J
Machens, Christian K
author_facet Calaim, Nuno
Dehmelt, Florian A
Gonçalves, Pedro J
Machens, Christian K
author_sort Calaim, Nuno
collection PubMed
description Neural systems are remarkably robust against various perturbations, a phenomenon that still requires a clear explanation. Here, we graphically illustrate how neural networks can become robust. We study spiking networks that generate low-dimensional representations, and we show that the neurons’ subthreshold voltages are confined to a convex region in a lower-dimensional voltage subspace, which we call a 'bounding box'. Any changes in network parameters (such as number of neurons, dimensionality of inputs, firing thresholds, synaptic weights, or transmission delays) can all be understood as deformations of this bounding box. Using these insights, we show that functionality is preserved as long as perturbations do not destroy the integrity of the bounding box. We suggest that the principles underlying robustness in these networks — low-dimensional representations, heterogeneity of tuning, and precise negative feedback — may be key to understanding the robustness of neural systems at the circuit level.
format Online
Article
Text
id pubmed-9307274
institution National Center for Biotechnology Information
language English
publishDate 2022
publisher eLife Sciences Publications, Ltd
record_format MEDLINE/PubMed
spelling pubmed-93072742022-07-23 The geometry of robustness in spiking neural networks Calaim, Nuno Dehmelt, Florian A Gonçalves, Pedro J Machens, Christian K eLife Computational and Systems Biology Neural systems are remarkably robust against various perturbations, a phenomenon that still requires a clear explanation. Here, we graphically illustrate how neural networks can become robust. We study spiking networks that generate low-dimensional representations, and we show that the neurons’ subthreshold voltages are confined to a convex region in a lower-dimensional voltage subspace, which we call a 'bounding box'. Any changes in network parameters (such as number of neurons, dimensionality of inputs, firing thresholds, synaptic weights, or transmission delays) can all be understood as deformations of this bounding box. Using these insights, we show that functionality is preserved as long as perturbations do not destroy the integrity of the bounding box. We suggest that the principles underlying robustness in these networks — low-dimensional representations, heterogeneity of tuning, and precise negative feedback — may be key to understanding the robustness of neural systems at the circuit level. eLife Sciences Publications, Ltd 2022-05-30 /pmc/articles/PMC9307274/ /pubmed/35635432 http://dx.doi.org/10.7554/eLife.73276 Text en © 2022, Calaim, Dehmelt, Gonçalves et al https://creativecommons.org/licenses/by/4.0/This article is distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/) , which permits unrestricted use and redistribution provided that the original author and source are credited.
spellingShingle Computational and Systems Biology
Calaim, Nuno
Dehmelt, Florian A
Gonçalves, Pedro J
Machens, Christian K
The geometry of robustness in spiking neural networks
title The geometry of robustness in spiking neural networks
title_full The geometry of robustness in spiking neural networks
title_fullStr The geometry of robustness in spiking neural networks
title_full_unstemmed The geometry of robustness in spiking neural networks
title_short The geometry of robustness in spiking neural networks
title_sort geometry of robustness in spiking neural networks
topic Computational and Systems Biology
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9307274/
https://www.ncbi.nlm.nih.gov/pubmed/35635432
http://dx.doi.org/10.7554/eLife.73276
work_keys_str_mv AT calaimnuno thegeometryofrobustnessinspikingneuralnetworks
AT dehmeltfloriana thegeometryofrobustnessinspikingneuralnetworks
AT goncalvespedroj thegeometryofrobustnessinspikingneuralnetworks
AT machenschristiank thegeometryofrobustnessinspikingneuralnetworks
AT calaimnuno geometryofrobustnessinspikingneuralnetworks
AT dehmeltfloriana geometryofrobustnessinspikingneuralnetworks
AT goncalvespedroj geometryofrobustnessinspikingneuralnetworks
AT machenschristiank geometryofrobustnessinspikingneuralnetworks