Philosophical Foundations

[Sergio Navega]

Nate Cull wrote:
>
>Sergio Navega wrote:
>
>> Categorization is difficult to do (and is important to AI) because it
>> can easily lead to intractability: an apple is a fruit, belong to the
>> category of edible stuff, is part of the vegetal kingdom. But it is
>> also an exemplar of round objects, solid objects, textured surfaces,
>> rotting things, etc. I could find thousands of "irrelevant" categories
>> in which an apple would fit. This "explosion" of categories, if not
>> taken care properly, would fill the memory of an agent rapidly,
>> exactly the opposite of what compression is supposed to do. 
>
>
>But presumably a good compressor would only generate a "category" when
>it was required in order to differentiate two otherwise similar
>objects?  And it would only differentiate the objects when there was
>some compelling environmental need to do so?  (Like, for example,
>needing to learn the difference between a banana and a wasp.  They're
>both yellowish, pointed things... but one stings and the other tastes
>good.)

I agree with your vision. 
One of the greatest problems that cognitive approaches to AI have to
face today is exactly what criteria must be used by the system to
do categorization. My idea is to let categorization occur as the
result of an external need to differentiate things that were
previously similar. For a normal adult, an apple is an apple. But
to a "connoisseur", you may have several "kinds" of apples, each
one with a different name, based on some characteristic (size,
sweetness, etc).

Regards,
Sergio Navega.