Weaknesses of this simulation
- We're interested in the way the brain would solve this task, and
we're a long way from that.
- One thing that is wrong is the vector input. Brain's use
senses.
- The Pearson measurement is also not legitimate. The brain doesn't
categorise by memorising firing behaviour and then comparing.
Pearson also requires you to run on the training set, in this
case radically increasing the simulation length.
- The training regime is not psychologically plausible. (750 ms.
per epoch.)
- There is no persistence and thus there are no cell assemblies.
