Hi Alex,
On 1/8/20 4:39 PM, Alexander Gabriel wrote:
and the following inference rule (because I couldn't find PLN rule that
covers this case):
The existing deduction rules
https://github.com/opencog/pln/blob/master/opencog/pln/rules/term/deduction.scm
are only defined for implication and inheritance links, so if you were
to use them you'd have to formulate the semantics of `linked` in these
terms. Let know if you want more guidance on that. That is said writing
your own rules is a perfectly fine way to go and in fact better if you
wish to have full control over the inference process.
When I now run the backward chainer on my KB with a query like:
|
query =EvaluationLink(
PredicateNode("linked"),
ListLink(ConceptNode("WayPoint001"),ConceptNode("WayPoint004")))
|
that works fine:
|
(SetLink(EvaluationLink(PredicateNode"linked")(ListLink(ConceptNode"WayPoint001")(ConceptNode"WayPoint004"))))
(stv 1.000000 0.000000)
|
The truth value (stv 1 0) however doesn't seem right.
but when I run it over a longer distance like this:
|
query = EvaluationLink(
PredicateNode("linked"),
ListLink(ConceptNode("WayPoint001"),ConceptNode("WayPoint005")))
|
the inference fails to produce the correct truth value:
|
(SetLink(EvaluationLink(PredicateNode"linked")(ListLink(ConceptNode"WayPoint001")(ConceptNode"WayPoint005"))))
(stv 1.000000 0.000000)
|
That answers looks like the previous one, with the likely wrong default TV.
I also seem to misunderstand how inference works because if I try to use
the EvaluationLink part (now with VariableNodes) in a GetLink query like
this:(so I can combine it with my other constraints), that only returns
pairs of Waypoint concepts that I have set explicitly, but none that
follow from the inference rule.
|
query =GetLink(
VariableList(
VariableNode("origin"),
VariableNode("destination")),
EvaluationLink(
PredicateNode("linked"),
ListLink(
VariableNode("origin"),
VariableNode("destination"))))
|
The results obtained by inference should normally be added to your
atomspace and thus the query should return them.
So something unintended seems to be happening. Can you point me to the
full script of your experiment?
Nil
Any help is highly appreciated.
Best,
Alex
--
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 [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/opencog/bb96ded1-42c4-8e6e-1502-608eeeb41fce%40gmail.com.
