Hi Shane & Pei,

>
> For instance, reasoning plays a central role in the system, but I haven't
> known what is new there. The basic operations include "forward inference,
> subgoaling, grounding, representing alternate worlds, and identity
> matching", which sounds strange to me --- they don't form a reasonable
> partition of reasoning.
>
> I'm sure Ben has more to say about Polyscheme vs. Novamente, since he
> talked
> with Cassimatis during the conference.
>

Ben just left for vacation, so I'll stop lurking and take a shot at
this... *cackles madly*

Polyscheme is firmly in the GOFAI tradition. The two things that make it
fairly unique, as far as I can tell, are the "procedural substrate", and
the "representational substrate".

Procedural substrate is the is the idea that all of the "AI" algorithms
can be implemented in terms of a few basic operations, basically local
manipulations of a candidate-solution. To quote from the page "Most
high-order reasoning and problem solving algorithms (even those from very
different subfields of AI, such as logic-theorem proving and probabilistic
inference) can be implemented using the same set of basic computational
operation". From my perspective (designing and implmenting the procedure
representation/learning component of Novamente, with a strong evolutionary
bias), something big is missing here; you can't really describe EC or
global probabalistic model-building within this framework; Ben and I spent
an annoying 15 minutes trying to talk to Nick about this, but really
weren't speaking the same language at all.

Representation substrate is the idea of mapping every problem into a
particular domain; physical reasoning. Arguably there is a lot of cogsci
that claims something like this for humans, but that's a separate issue;
suffice it to say that this idea is not very Novamnete-ish at all; the
Novamente philosophy is more along the lines of "find some domain to
reason *effectively* about any given problem", which may well be physical
reasoning, for some problems (those involving phsyical reasoning in
particular ;->), but certainly not all of them.

Regards,
Moshe

-------
To unsubscribe, change your address, or temporarily deactivate your 
subscription, 
please go to http://v2.listbox.com/member/[EMAIL PROTECTED]

Reply via email to