Hebbian Learning
- The only types of learning for which there is neural evidence
is Hebbian learning, cell creation and, cell death
- In the broad sense, Hebbian learning is change in synaptic
weight based on activation of the pre and post-synaptic neurons
- If neuron 1 and 2 fire and they are connected, the connection
is strengthened
- The anti-Hebbian rule states that if one of those neurons fire, and
the other doesn't the connection is weakened
- The problem happens when a bunch of neurons are not in CAs and there
is spontaneous activation
- Usually, only one neuron around a synapse will be active
- The anti-Hebbian rule will be applied a lot more than the
Hebbian rule
- The synaptic weights will go to zero
- Similar rules apply to inhibitory neurons
- The inhibitory values will go to negative infinity
