Conclusion
- CABot1 is a functioning agent all in neurons.
- CABot2 is a bit better but has a symbolic hack (that will
be sorted for CABot3).
- They don't do much learning, but
- they are functioning neural (not connectionist) agents.
- CABot3 should be even better with texture recognition in
vision, and spatial cognitive mapping.
- Future work will improve learning in the system to enable
it to learn more from its environment.
- If we find problems we can't solve, we'll look at how
humans do it.
- If someone points out that we solve the problem in a manner
different from humans, we'll check, but then change.
