Brain Markov
Markov Brains are collections of gates (which in turn are mini-brains). Usually, in a single Markov Brain update, all of the brains gates are run in parallel; Gates read from Nodes and write to Nodes Next.
Gates have inputs and outputs which connect to brain nodes. Like Brains, gates read inputs, process the inputs and produce outputs. Gates can have arbitrary numbers of inputs and outputs, which will be defined by the gate type.
Markov Brains are collections of gates. The gates within a single brain may be of different types. User defined gates can be added trivially.
- Deterministic Gate
- uses an internal look up table to translate binary inputs to binary outputs.
- Probabilistic Gate
- uses an internal look up table to convert binary inputs to semi random binary outputs (based on internal probabilities).
- Trit Deterministic Gate
- like deterministic gate, but inputs; and outputs can also be -1
- Genetic Programming Gate
- preforms math operations on inputs (+ or - continuous value) to generate outputs (+ or - continuous value)
- Threshold Gates
- accumulates binary inputs until threshold is reached, than delivers 1 to output.
- Neuron Gates (more complex threshold gate)
- accumulates input values (charge) until threshold (+ or - continuous value) is reached, than deliver output value (+ or - continuous value).
- Void Gate
- Like Deterministic gates, but includes a user defined error rate. When an error occurs, one output is set to 0 (this may result in the original output).
- Fixed Epsilon Gates
- Like Deterministic gates, but includes a user defined error rate. When an error occurs, the input value is changed so a different set out outputs is delivered from the gates internal lookup table.
see settings file.
Markov Brains inherit the variables from Brains and have the following additional variables
- gates - list of gates in this brain
- GLB - a Gate List Builder (builds the gates list - e.g. converts a genome into gates)