Saturday, June 30, 2007, Vladimir Nesov wrote:

VN> But diversity of challenges in reasoning itself seems to be greater
VN> than what might be necessary for NLP. So why start at conceptual level?
VN> Internal KR language can be used occasionally to steer pattern
VN> formation and observe reasoning internals, and natural language can
VN> be used for most of the teaching, from the very start.

Btw, this analogy just came to me: there's this AGISim thing going for
Novamente. Are all meaningful concepts created within the system and
utilized for its reasoning have each been manually provided be teacher
as atomic tags? Or it just has a bunch of modality-provided tags, like 'left',
'right', 'big', 'box' and comes up by itself with complex situation
concepts (like 'fetch'), subsequently used as atomic entities in
multiple situations?

Teching anything complex looks difficult
without a language (character-based or otherwise) - no way to
describe reasoning, hence it must learn much all by itself...

-- 
 Vladimir Nesov                            mailto:[EMAIL PROTECTED]

-----
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/?member_id=231415&user_secret=e9e40a7e

Reply via email to