Cargando…
Deep neural networks using a single neuron: folded-in-time architecture using feedback-modulated delay loops
Deep neural networks are among the most widely applied machine learning tools showing outstanding performance in a broad range of tasks. We present a method for folding a deep neural network of arbitrary size into a single neuron with multiple time-delayed feedback loops. This single-neuron deep neu...
Autores principales: | Stelzer, Florian, Röhm, André, Vicente, Raul, Fischer, Ingo, Yanchuk, Serhiy |
---|---|
Formato: | Online Artículo Texto |
Lenguaje: | English |
Publicado: |
Nature Publishing Group UK
2021
|
Materias: | |
Acceso en línea: | https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8397757/ https://www.ncbi.nlm.nih.gov/pubmed/34453053 http://dx.doi.org/10.1038/s41467-021-25427-4 |
Ejemplares similares
-
Editorial: From Structure to Function in Neuronal Networks: Effects of Adaptation, Time-Delays, and Noise
por: Cabral, Joana, et al.
Publicado: (2022) -
Delay-induced patterns in a two-dimensional lattice of coupled oscillators
por: Kantner, Markus, et al.
Publicado: (2015) -
Consequences of delays and imperfect implementation of isolation in epidemic control
por: Young, Lai-Sang, et al.
Publicado: (2019) -
Autonomous epithelial folding induced by an intracellular mechano–polarity feedback loop
por: Wen, Fu-Lai, et al.
Publicado: (2021) -
Self-organized noise resistance of oscillatory neural networks with spike timing-dependent plasticity
por: Popovych, Oleksandr V., et al.
Publicado: (2013)