Hi,

> In a private email, Moshe Looks added a common complaint that I'd
> forgotten:
>
> ----
>
> Complaint: Learning to be intelligent isn�t possible without building up
> abstract cognitions hierarchically from a foundation of content-rich
> sensory
> and action streams.   While Novamente can deal with content-rich sensory
> and
> action streams in principle, it�s not really centrally designed for this.
> Work on AGI should begin with study of perception and action, and then one
> should ask what sort of cognition naturally goes along with one�s working
> perception and action modules � and the answer may or may not look like
> Novamente.
>
Well, this is a bit of a straw-man. I guess its fairly close to the view
of Patrick Winston and Project Genesis (http://genesis.csail.mit.edu/),
but *I* was applying the words "content-rich" to modalities rather that
sensory and action streams. For example, would you condsider a 30x30
binary pixel grid with a logo turtle "content-rich sensory and action
streams"? Maybe, but I'm sure its not the first image that appears in the
minds of the intrepid AGIers reading this ;-). Yet you can come up with
all sorts of "rich" modality-specific  problems, like resolving ambiguity,
object completion, invariance under transformations, temporal patterns,
etc, etc, etc... Clearly a buch of the brain's perception/action
complexity is completely irrelevant in dealing with such a world, but not
all of it! The Anwser still bascially works, I just wanted to get the
question clear..

Moshe

****************************

> Answer: Our intuition is that content-rich media aren�t critical, rather
> that what�s important for learning to think is interaction with other
> minds
> in a shared perceptual environment in which you�re embodied.  However, if
> content-rich media are critical, Novamente can be used for rich
> sensorimotor
> processing perfectly well.  While it�s always possible to code specialized
> processing code for each type of sensor and actuator, we believe it�s
> better
> to begin with a common framework (such as BOA+PTL) and then specialize it
> to
> deal with the different modalities.  This is conceptually analogous to the
> way the brain uses the same basic neural mechanisms to deal with the
> different human modalities, and also with cognition.
>
> ----
>
>> Complaint: The design is too complicated, there are too many
>> parts to coordinate, too many things that could go wrong
>>
>> Answer: Yes it IS complicated, and we wish it were simpler, but
>> we haven�t found a simpler design that doesn�t seem patently
>> unworkable.  Note that the human brain is also mighty complicated
>> � this may just be the nature of making general intelligence work
>> with limited resources.
>>
>> Complaint: BOA and PTL are not enough, you need some kind of more
>> fundamentally innovative, efficient, or (whatever) learning
>> algorithm.  This complaint never comes along with any suggestion
>> regarding what this �mystery algorithm� might be, though � most
>> often it is hypothesized that detailed understanding of the human
>> brain will reveal it.
>>
>> Answer: This is possible, but it seems to us that a hybrid of BOA
>> and PTL will be enough.  The question is whether deeper
>> integration of BOA and PTL than we�ve done now will allow BOA
>> learning of reasonably large (500-1000 node) combinator trees.
>> If so, then we almost surely don�t need any other learning
>> algorithm, though other algorithms may be helpful.
>>
>> Complaint: You�re programming in too much stuff: you should be
>> making more of a pure self-organizing learning system without so
>> many in-built rules and heuristics
>>
>> Answer: Well, the human brain seems to have a lot of stuff
>> programmed in, as well as a robust capability for self-organizing
>> learning.  Conceptually, we love the idea of a pure
>> self-organizing learning system as much as anyone, but it doesn�t
>> seem to be feasible given realistic time and processing power and
>> memory constraints.
>>
>> Complaint: Programming explicit logical rules is just wrong;
>> logic should occur as an emergent phenomenon from more
>> fundamental subsymbolic dynamics
>>
>> Answer: Probabilistic logic is not necessarily symbolic; in the
>> Novamente design we use PTL for both subsymbolic and symbolic
>> learning, which we believe is a highly elegant approach.  The
>> differences between subsymbolic probabilistic logic and e.g.
>> Hebbian learning are not really very great when you look at them
>> mathematically rather than in terms of verbiage.  The Novamente
>> design is not tied to programming-in logical knowledge a la Cyc.
>> It�s true that the PTL rules are programmed in (though in
>> Novamente 2.0 they will be made adaptable), but this isn�t so
>> different from the brain having particular kinds of long-term
>> potentiation wired in, is it?
>
>
> -------
> To unsubscribe, change your address, or temporarily deactivate your
> subscription,
> please go to http://v2.listbox.com/member/[EMAIL PROTECTED]
>

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

Reply via email to