gts wrote:

> I'm not expecting "essentially perfect" coherency in AGI.
> I understand perfection is out of reach.

My question to you was whether, as a professed C++ developer, you are
familiar with the well-known impracticality of certifying a non-trivial
software product to be essentially free of unexpected failure modes, and
if so, do you see a similarity to your question of coherent reasoning by
machine intelligence?

In a similar vein, do you think you understand Ben's comment about the
problem being NP-hard?
 
 
> However, as I was explaining, accuracy in one's probabilistic
> judgements is not the same as coherent reasoning about one's
> probabilistic judgements. One can arrive at terribly wrong
> conclusions without being incoherent in the sense meant by 
> De Finetti.

Yes, coherence does not imply truth, due to inherently limited context.

By the way, De Finetti used the word "coherent" in the very standard
sense meaning that all the pieces must fit together from all possible
points of view (within all possible contexts.)

This same concept of coherence is the basis of the axioms of probability
and the principle of indifference.

Understand this underlying concept and you may understand the others.


> What is a machine intelligence if it is not a machine capable of
coherent
> reasoning? 

You keep saying this, but then just above above you state that you
realize "essentially perfect coherency...is out of reach." Are you
waffling here, or just unclear?
 
You keep referring to De Finetti's Dutch Book argument, which does ask
us to consider perfect coherence to make its perfectly valid point, but
never claims such is achievable in practice.  We can understand what it
means to always refuse a Dutch Book bet, but we will forever lack the
omniscience to always know and act in such a coherent way.

A machine intelligence will be able to reason more coherently than a
human because it will be able to reason within greater context, and more
importantly, it can be designed to be free of some inappropriate human
biases, but this is vastly different from saying that it is could
"reason coherently" in any absolute sense.

Coherence. You keep using that word. I do not think it means what you
think it means.

- Jef

-----
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

Reply via email to