On 10/21/2016 08:44 AM, Ben Goertzel wrote:
Yes, those 2 points would satisfy our current needs... thx...
Looking at the code, I'm thinking it's gonna be simpler to create a new tv type, possibly inheriting from SimpleTV.
That is because the base TV class is completely abstract, adding the methods is only gonna force other sub-classes to implement this, which is a pain, although might be what we want, but since its usage is not entirely sure to me at that point I suppose we should go with the simpler stuff first.
So I intend to create an EvidenceCountTruthValue (possibly inheriting from SimpleTruthValue) + scheme binding. That should cover your current needs. Let me know otherwise.
Nil
On Oct 21, 2016 1:23 PM, "Nil Geisweiller" <[email protected] <mailto:[email protected]>> wrote: Hi, I think someone suggested a while ago that positive and total counts be the main data stored in a tv, and the probability would be calculated on the fly when requested. The API could offer methods to access and modify just the positive count, and so if the total count is let undefined the probability would be NAN. This could be done in a way that maintains compatibility with the existing TV API, so there would be no need to introduce another TV type. Not that introducing a new TV type is bad, but I feel it is both sufficiently simple and useful across types that we might just put it in the base class. What do you think? Regarding the total count, I see what you mean, on an abstract level at least. So let's put the requirement: whenever an reasoning rule can be agnostic of the total count and some if its premises have undefined total count, the rule should be able to propagate just the positive count. Do these 2 points above entirely capture your needs? On 10/20/2016 05:57 PM, Ben Goertzel wrote: Whereas, the InheritanceLink after M2I means "X is a member of mouse, implies (extensionally and intensionally) X is a member of the set of things that cats eat" Yep, that's what I read too. Hmm, this 2 layers (elements of elements) inclusion gives me the creep. ;-) But I'm not sure why, hopefully experimentation will help us to get this straight, and this is coming soon. Nil I.e. the M2I rule in its currently implemented form relies on the interpretation of ConceptNode "cat" as equivalent to SatisfyingSet MemberLink $X ConceptNode "cat" No, it is the count of (SetLink (Concept "Ben")), which is 1. Or if that weird M2I as currently implemented rule is correct, it is the count of It's the count of ConceptNode "cat" which is the count of SatisfyingSet MemberLink $X ConceptNode "cat" Agreed. Ultimately I think we could offer a positive_evidence setter. But I still don't understand the need for propagating positive_evidence without knowing the total count. It's because the same positive-evidence value makes sense in the context of multiple different choices for the total count... -- Ben
-- 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 post to this group, send email to [email protected]. 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/5809D09F.2030808%40gmail.com. For more options, visit https://groups.google.com/d/optout.
