Hi,
I wonder if the order of arguments in intensional inheritance shouldn't
be swapped. Let me recall the definitions of the various inheritances
(according to the PLN book [and coincidingly the opencog wiki])
1. Extensional Inheritance
ExtInh A B
is equivalent to
SubSet A B
2. Intensional Inheritance
IntInh A B
is equivalent to
ExtInh A_{PAT} B_{PAT}
where A_{PAT} is the (fuzzy) set of patterns of A (i.e. super sets of A,
that have discriminative power and low complexity).
3. Mixed Inheritance
Inh A B
is equivalent to
Or
ExtInh A B
IntInh A B
Here's the problem, if A intensional inherits from B, then it means that
A tends to have less patterns than B, meaning A is as abstract as B or
more, or equivalently B is as specialized as A or more.
However, in average the more something is specialized, the smaller its
extension will be. Thus it is expected that
ExtInh A B
will tend to be correlated with
IntInh B A
and anti-correlated with
IntInh A B
For that reason I think either the definition of mixed inheritance
should be redefined into
Or
ExtInh A B
IntInh B A
or the definition of intensional inheritance should be redefined into
ExtInh B_{PAT} A_{PAT}
Otherwise mixed inheritance will tend to be always half true, in average
regardless of A and B, because of the anti-correlation tendency between
ExtInh A B and IntInh A B.
Nil
--
You received this message because you are subscribed to the Google Groups
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to opencog+unsubscr...@googlegroups.com.
To post to this group, send email to opencog@googlegroups.com.
Visit this group at https://groups.google.com/group/opencog.
To view this discussion on the web visit
https://groups.google.com/d/msgid/opencog/2cdf3d5b-fb0e-ac2a-c5d6-72c4a7cff0ba%40gmail.com.
For more options, visit https://groups.google.com/d/optout.
