Matt,
What are you being so tetchy about? The issue is what it takes for any
agent, human or machine.to understand information .
You give me an extremely complicated and ultimately weird test/paper, which
presupposes that machines, humans and everyone else can only exhibit, and be
tested on, their thinking and understanding in an essentially Chinese room,
insulated from the world.
I am questioning, and refuting the entire assumption, behind those
extraordinarily woolly ideas of Turing, (witness the endlessly convoluted
discussions of his test on this group - which clearly people had great
difficulty "understanding" precisely because it is so woolly, when you try
to understand exactly what's testing).
An agent understands information and information objects,IMO, if he can
point to the real objects referred to in the real world, OUTSIDE any
insulated room. (I am taking Searle one step further). It is on his ability
to use language to engage with the real world, - fulfil commands/requests
like where's the key?, "what food is in the fridge?" "is the room tidy?"
(and progressively more general information objects), that an agent's
understanding must be tested.
That is consistent with every principle that you seem to like to invoke, of
evolutionary fitness. Language and other forms of information exist
primarily to enable humans to deal with real objects - and to survive - in
the real world, and not in any virtual world, that academics and AGI-ers
prefer to inhabit.
My special distinction, I think, is v. useful - the Chinese translator and
AGI's "comprehend" information/language - merely substituting symbols for
other symbols. The agent who can use that language to deal with real
objects, truly *understands* it.
This explanation is consistent with how humans actually fail to understand
on inumerable occasions, and also how computers and would-be AGI's fail to
understand - not just outside in the real world, but *inside* their
rooms/virtual worlds. All language understanding collapses without real
object/world engagement.
In case you are unaware how academics will go to quite extraordinary mental
lengths to stay inside their rooms, see this famous passage which helped
give birth to science , - re natural philosophers who, (with small
modifications, like AGI-ers)
"having sharp and strong wits, and abundance of leisure, . as their persons
were shut up in the cells of monasteries and colleges, and knowing little
history, either of nature or time, did out of no great quantity of matter,
and infinite agitation of wit spin out unto those laborious webs of learning
which are extant in their books. For the wit and mind of man, if it work
upon matter, worketh according to the stuff; but if it work upon itself, as
the spider worketh his web, then it is endless, and brings forth indeed
cobwebs of learning, admirable for the fineness of thread and work, but of
no substance or profit." Francis Bacon, The Advancement of Learning.
.
Matt:
"To understand/realise" is to be distinguished
from (I would argue) "to comprehend" statements.
How long are we going to go round and round with this? How do you know if
a machine "comprehends" something?
Turing explained why he ducked the question in 1950. Because you really
can't tell. http://www.loebner.net/Prizef/TuringArticle.html
-- Matt Mahoney, [EMAIL PROTECTED]
-------------------------------------------
agi
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/
Modify Your Subscription:
https://www.listbox.com/member/?&;
Powered by Listbox: http://www.listbox.com
-------------------------------------------
agi
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/
Modify Your Subscription:
https://www.listbox.com/member/?member_id=8660244&id_secret=114414975-3c8e69
Powered by Listbox: http://www.listbox.com
