This is an interesting thread, I'll add some comments:
1. For KR purposes, I think first order predicate logic is a good choice. Geniform 2.0 can be expressed in FOL entirely. ANN is simply not in a state advanced enough to represent complex knowledge (eg things that are close to NL). I also agree that ANN is suitable for low-level vision processing, but as for high level cognition I don't see any distinct advantage that it may offer that logic does not possess. This is a big issue and I'll post my answer on my FAQ page.
2. Ben raised the issue of learning. I think we should divide learning into 3 parts:
(1) linguistic eg grammar
(2) semantic / concepts
(3) generic / factual.
I think the learning mechanisms in (1)-(3) are quite distinct, for example, learning that "fish cat eat" is ungrammatical, and learning that cats like to eat fish, are quite distinct matters.
I have somewhat addressed (2) and (3) on my web page. I have not looked at language learning yet.
As for factual learning, I think the most important mechanism is *consistency seeking*, which is a general form of belief revision. The baby AI's mind contains a whole lot of random facts, and as the mind matures, the web of facts get more and more consistent. Facts may support each other within the knowledgebase, for example "all men are lazy" is supported by "John is often lazy", etc.
When internal inconsistency is discovered (eg upon new facts arrival), the KB decides which facts should be retracted or modified, based on the amount of "support" from other facts. A fact that contradicts many other existing facts would be regarded as "ridiculous". (Maybe that's why enlightening stupid people is so frustrating, because their minds contain a lot of bad quality "facts").
That'd be my unsupervised algorithm for factual learning.
As for concept learning it may also be based on consistency seeking. Another mechanism is inductive logic programming, which is really a coarse-grained version of feature-space clustering.
YKY
This list is sponsored by AGIRI: http://www.agiri.org/email To unsubscribe or change your options, please go to: http://v2.listbox.com/member/?list_id=303
