Output value
Neuron's bias here
delta bias previous
delta bias total
derivative state
error gated
error projected
error responsibility
Connections this node gates
Node ID for NEAT
Incoming connections to this node
index
Used for dropout. This is either 0 (ignored) or 1 (included) during training and is used to avoid overfit.
Outgoing connections from this node
previous state
A self connection
state
The type of this node.
Actives the node.
When a neuron activates, it computes its state from all its input connections and 'squashes' it using its activation function, and returns the output (activation).
You can also provide the activation (a float between 0 and 1) as a parameter, which is useful for neurons in the input layer.
A neuron's 'Squashed' output value
This node gates (influences) the given connection
Connection to be gated (influenced) by a neuron
Clears this node's state information - i.e. resets node and its connections to "factory settings"
node.clear() is useful for predicting time series.
Connects this node to the given node(s)
Node(s) to project connection(s) to
Initial connection(s) weight
If true connect nodes to each other
Disconnects this node from the given node(s)
Node(s) to remove connection(s) to
Is this a hidden Node?
Is this a input Node?
Is this a output Node?
Checks if the given node(s) are have outgoing connections to this node
Checks if node(s) have outgoing connections into this node
Returns true, if every node(s) has an outgoing connection into this node
Checks if this node has an outgoing connection(s) into the given node(s)
Checks if this node has outgoing connection(s) into node(s)
Returns true, if this node has an outgoing connection into every node(s)
Mutates the node's activation function
Mutates the node's bias
The method is needed for the min and max value of the node's bias otherwise a range of [-1,1] is chosen
Backpropagate the error (a.k.a. learn).
After an activation, you can teach the node what should have been the correct output (a.k.a. train). This is done by backpropagating. Momentum adds a fraction of the previous weight update to the current one. When the gradient keeps pointing in the same direction, this will increase the size of the steps taken towards the minimum.
If you combine a high learning rate with a lot of momentum, you will rush past the minimum (of the error function) with huge steps. It is therefore often necessary to reduce the global learning rate ยต when using a lot of momentum (m close to 1).
The target value (i.e. "the value the network SHOULD have given")
More options for propagation
Momentum adds a fraction of the previous weight update to the current one.
When set to false weights won't update, but when set to true after being false the last propagation will include the delta weights of the first "update:false" propagations too.
Stops this node from gating (manipulating) the given connection(s)
Connections to remove gate - i.e. remove this node from
Set activation type
the new activation type
Set bias.
the new bias value
Converts the node to a json object that can later be converted back
A node representing json object
Generated using TypeDoc
Creates a new neuron/node
Neurons are the basic unit of the neural network. They can be connected together, or used to gate connections between other neurons. A Neuron can perform basically 4 operations: form connections, gate connections, activate and propagate.
For more information check: