- One of the big problems with neural systems is that binding
- In symbolic VonNeumann systems, variables just get assigned
values. i := 1;
- It's hard to bind with neurons; the red square blue circle problem.
- This system does it with short-term potentiation (STP).
- So, via Hebbian learning, the firing neurons become bound.
- The slot is filled with the value.
- The binding is slowly erased (that's the short) automatically.