Difference between revisions of "SensorsDigital2NeuralTransformation"

From aHuman Wiki
Jump to: navigation, search
(Automated page entry using MWPush.pl)
(No difference)

Revision as of 01:29, 16 June 2015

Converting specific external stream to neural stream

@@Home -> SensorsResearch -> SensorsDigital2NeuralTransformation


Capture events

Certain:

  • conscious capture is performed by event
  • event is either control change command or spotted changes in external properties
  • conscious capture depends on capture resolution and capture area

Uncertain:

  • sensor has unconscious ongoing capture with constant low resolution but wide area
  • sensor can generate either full set of properties or only delta to previous transfer

External stream

Certain:

  • one external stream item is transformed into zero, one or more sensor stream item
  • external item can be ignored if sensor stream is not consumed in time

Uncertain:

  • set of events produce external stream

Sensor stream

Certain:

  • sensor stream has neural representation
  • sensor stream item is represented by vector of variables
  • sensor stream item is represented by sequence of vectors
  • sequence can be of different length depending on focus
  • width is the number of vector variables
  • all items of one sequence have the same width - sequence width

Uncertain:

  • vector variable is float
  • sensor stream is stateless - every sensor item is derived only from one external item
  • different sequences can have different width
  • sequence has energy which is proportional to sequence width and length
  • sequence energy is external stream item energy multiplied by energy amplifier
  • energy amplifier factor is part of focus conrol state

Sensor neuron

Certain:

  • each vector variable corresponds to specific sensor neuron
  • sensor neuron can have multiple connections to neocortex networks
  • one sensor neuron represents the same external properties for the same resolution

Uncertain:

  • sensor neuron does not have fixed semantic
  • one sensor neuron can represent different external properties for different resolutions

Finally for the project purposes

  • conscious capture by event
  • full property set transfer
  • throtteling of external stream
  • external stream item is represented by variable-length sequence of sensor stream items
  • float variables