References & Citations
Computer Science > Neural and Evolutionary Computing
Title: Toward Neuromic Computing: Neurons as Autoencoders
(Submitted on 4 Mar 2024 (v1), last revised 25 Apr 2024 (this version, v4))
Abstract: This short paper presents the idea that neural backpropagation is using dendritic processing to enable individual neurons to perform autoencoding. Using a very simple connection weight search heuristic and artificial neural network model, the effects of interleaving autoencoding for each neuron in a hidden layer of a feedforward network are explored. This is contrasted to the standard layered approach to autoencoding. It is shown that such individualised processing is not detrimental and can improve network learning.
Submission history
From: Larry Bull [view email][v1] Mon, 4 Mar 2024 18:58:09 GMT (532kb)
[v2] Wed, 6 Mar 2024 17:49:56 GMT (582kb)
[v3] Mon, 11 Mar 2024 09:55:33 GMT (630kb)
[v4] Thu, 25 Apr 2024 16:01:29 GMT (734kb)
Link back to: arXiv, form interface, contact.