Persistence
- When a neuron fires, it loses all its activation, so you need
bunches of neurons (connected appropriately) to get them to
persistently fire.
- The neuropsychology of it all calls these things Cell Assemblies (CAs).
- People (rats and macaques) have CAs for concepts. So I have a
dog CA.
- A really simple model of this is just a well connected set of
neurons (and there are real short comings too).
- So, you fire all the neurons once, and they continue to fire.
- In our FLIF model, they can become fatigued after a certain time
and that helps with one of short comings.
- So when a word is presented, this activates neurons that
fire persistently.
- These activate the appropriate lexical and semantic neurons.
- Rules similarly get activated, and when the rule is done being
applied, it inhibits the words, (and gets the next word).
