Neural Cognitive Models
- I started with an ax to grind; Smolensky (1988) says
"neural models of cognitive processes are ... currently
not feasible".
- I gather Roman presented our neural model of a two choice task.
- I probably should review some other neural models here.
- So, you clearly can build neural cognitive models.
- Neurons are Turing complete.
- So, it does leave open the question of how neurally accurate a
neural cognitive model is.
- None the less, it's nice to see you can do it, and simulated
models can be increasingly improved by matching more constraints.
- They can also make use of direct evidence.
