Some early work on JOONN
InputMatrix - handles connecting data into the neural cube. typically at a front layer. Runs with its own scheduled thread
OutputMatrix - handles outputs with different paradigms - FileLogger, Alerter, CubeBridge
Now let's build up a Cube from small to large
Neuron
This is the base type and we will have several advanced types
uses:
Synapse
hasa Weight
hasa Value
hasa ThresholdFunction, DecayFunction
hasa Algorithm
hasa SpikingFunction
hasa InterconnectGrowth function (for creating new synaptic connections)
hasa Chaos function (to add gradual chaos into the system)
hasa PerformanceStats (how fast is it processing)
NeuralCore
hasa Value
hasa 64 value MemoryMap
hasa PassingMemoryMap for propagations
hasa MemoryMapTransferClass for constructing the outputMap
hasa SpikeInput, SpikeOutput
hasa ConnectionArchitecture (for what other neurons it links to)
hasa ThresholdFunction, DecayFunction
hasa PropagationFunction
Spike
This is used to coordinate between layers of analysis
hasa ReceiveMatrix
hasa DistributeMatrix
hasa SpikeThreshold
hasa value
hasa SpikingFunctino
Layer
One 2d layer of the cube
hasa height, width
hasa InterconnectModel (soyou dont have to hand wire up thousands of neurons!)
hasa zIndex
hasa neuralMatrix (x,y)
hasa NeuralType -- initially one class of neuron per layer is themaximumdiversity
NeuralCube
hasa height, width
hasa LayerList
hasa SpikeArray
hasa InputMatrix, OutputMatrix
HolographicValueMap
-- This is like a complex data store that is used for advanced recognition or memory. It unifies partial maps stored in neurons and other HVMs
No comments:
Post a Comment