spark.nn.components.learning_rules.zenke_rule

Classes

ZenkeRuleConfig	ZenkeRule configuration class.
ZenkeRule	Zenke plasticy rule model. This model is an extension of the classic Hebbian Rule.

Module Contents

class spark.nn.components.learning_rules.zenke_rule.ZenkeRuleConfig(__skip_validation__=False, **kwargs)

Bases: spark.nn.components.learning_rules.base.LearningRuleConfig

ZenkeRule configuration class.

Parameters:

__skip_validation__ (bool)

pre_tau: float | jax.Array | spark.nn.initializers.base.Initializer

post_tau: float | jax.Array | spark.nn.initializers.base.Initializer

post_slow_tau: float | jax.Array | spark.nn.initializers.base.Initializer

target_tau: float | jax.Array | spark.nn.initializers.base.Initializer

a: float | jax.Array

b: float | jax.Array

c: float | jax.Array

d: float | jax.Array

p: float | jax.Array

eta: float | jax.Array

class spark.nn.components.learning_rules.zenke_rule.ZenkeRule(config=None, **kwargs)

Bases: spark.nn.components.learning_rules.base.LearningRule

Zenke plasticy rule model. This model is an extension of the classic Hebbian Rule.

Init:

pre_tau: float | jax.Array post_tau: float | jax.Array post_slow_tau: float | jax.Array target_tau: float | jax.Array a: float | jax.Array b: float | jax.Array c: float | jax.Array d: float | jax.Array P: float | jax.Array eta: float | jax.Array

Input:

pre_spikes: SpikeArray post_spikes: SpikeArray kernel: FloatArray

Output:

kernel: FloatArray

Parameters:

config (ZenkeRuleConfig | None)

config: ZenkeRuleConfig

build(input_specs)

Build method.

Parameters:

input_specs (dict[str, spark.core.specs.PortSpecs])

reset()

Resets component state.

Return type:

None

__call__(pre_spikes, post_spikes, kernel)

Computes and returns the next kernel update.

Parameters:

• pre_spikes (spark.core.payloads.SpikeArray)

• post_spikes (spark.core.payloads.SpikeArray)

• kernel (spark.core.payloads.FloatArray)

Return type:

spark.nn.components.learning_rules.base.LearningRuleOutput