Local homeostatic regulation of the spectral radius of echo-state networks
Fabian Schubert, Claudius Gros
This paper introduces two local synaptic scaling rules for echo-state networks to control the spectral radius of recurrent synaptic weights, showing that flow control outperforms variance control in maintaining network performance under varying input conditions.
Frontiers in Computational Neuroscience · Feb 24, 2021Peer-reviewedNeuroscience