Key Word Search
by term...

by definition...

for letter: "R"

Ramachandran, Vilayanur S.
Random Number
Reactive Feedback
Receptor Pad
Recurrent Neural Network
Ref 1
Reinforcement Learning
Retention Time
Retroactive Interference
Reverse Transcription
Ribonucleic Acid
Richardson, Jane S.

Unless otherwise indicated, all glossary content is:
(C) Copyright 2008-2015
Dominic John Repici
No part of this Content may be copied without the express written permision of Dominic John Repici.


Receptor Pad

RP, also Dynamic Connection (DC) - In Netlab, a Receptor Pad (RP) is a site that is defined on a target (post-synaptic) neuron, where a connection (a synapse) with an axon may form. The axon forming the connection may be from ANY source and can form at the RP as the network runs. When a synapse has formed a connection at an RP, the synapse will remain between the source axon and post-synaptic cell at that RP unless the weight value of that synapse falls to zero or very near zero (for example). At that point the connection between the specific axon and the target will be released and the RP will be free to connect with a different axon (or the same one again).

A given neuron's RP's may be limited to forming connections with:
  • Only axons from other neurons (Intra-module).
  • Only axons from outside the current unit (Input-terminals)
  • Only axons from neurons that release specifically named chemical influences (CIs)

See the entry for Latent Connections for a discussion of the differences between LCs and RPs.

RPs are a Netlab abstraction. They are part of the function provided by Netlab to mimic the mechanisms of growth-cones and pathfinding observed in biological neural networks. RPs are essentially places where Netlab's growth-cone mechanisms can form synapses dynamically (where the pre- and post-synaptic units are decided at run-time, as opposed to already being defined at design-time).

Also: Latent Connection        


Web-based glossary software: (c) Creativyst, 2001-2015