Nil, not sure what you meant by low level either. The way I first understood > it is whether it should be built-in or emergent. I suppose that in > principle it could be emergent, but in practice it would dramatically > postpone the moment some general form of reasoning can take place, thus > it is handy to have it built-in. And as Ben said, it is only hardwired > at some places for efficiency reason. PLN rules being atoms it makes it > ultimately amenable to self-improvements, or to support other logics, > which is already the case, albeit somewhat brittly. >
I think you understood correctly. I thought PLN is entirely programmed in C++ layer, not in URE layer, so I got a bit confused. In my opinion, it is a good approach to have a general meta rule engine (URE) at the bottom and to hold all the knowledge in it, including PLN. And if something deserves to be faster, it is reasonable to make a faster copy in C++ or a similar language. However, keeping an URE copy of accelerated knowledge provides us a meta dimension to mechanically reason about it. Please, your feedback and criticisms are more than welcome. I like to > believe we're free-speech absolutists here and not easily offendable > (oops, I hope it's not gonna trigger a torrent of cursing). ;-) > Thank you very much, I like this community around here. And I don't believe in cursing, cursing pollutes thinking process, and forms intellectual extremes we should avoid. All knowledge should be tagged for possible improvements, and cursing makes improvements impossible. Like in that popular story: the more we know, the less certain we should be about the knowledge correctness :-) - Ivan V. - pon, 10. pro 2018. u 08:30 'Nil Geisweiller' via opencog < [email protected]> napisao je: > Ivan, > > not sure what you meant by low level either. The way I first understood > it is whether it should be built-in or emergent. I suppose that in > principle it could be emergent, but in practice it would dramatically > postpone the moment some general form of reasoning can take place, thus > it is handy to have it built-in. And as Ben said, it is only hardwired > at some places for efficiency reason. PLN rules being atoms it makes it > ultimately amenable to self-improvements, or to support other logics, > which is already the case, albeit somewhat brittly. > > On 12/9/18 7:55 PM, Ivan Vodišek wrote: > > for a long time now. So I kindly ask to excuse my, sometimes silly > > remarks, as I hope we could contribute each other's work. > > Please, your feedback and criticisms are more than welcome. I like to > believe we're free-speech absolutists here and not easily offendable > (oops, I hope it's not gonna trigger a torrent of cursing). ;-) > > Nil > > > > > Sincerely, > > - Ivan V. - > > > > ned, 9. pro 2018. u 18:00 Ben Goertzel <[email protected] > > <mailto:[email protected]>> napisao je: > > > > On Sun, Dec 9, 2018 at 11:37 AM Ivan Vodišek <[email protected] > > <mailto:[email protected]>> wrote: > > ... > > > I don't know, I never understood why is PLN defined at low level > > in OpenCog. Sure it's a cool asset, to know probability and > > confidence of any formula (not to degrade it, PLN surely does have a > > deal of scientific value), but why low level if AtomSpace is Turing > > complete. Starting from observation of natural general intelligence > > in humans, we don't have that specific probability feedback when we > > are thinking about something. But that doesn't mean we can't > > calculate it on our own, once we receive an interest stimulus about > > probability. > > > > > > > I'm not sure at what sense PLN is "low level" in the software sense? > > Software-wise PLN is one among many rule-systems definable using > > Atomspace and executable using URE rule engine... it doesn't have a > > distinguished position. The special truth-values used by PLN are a > > special case of general Values associated with Atoms etc. > > > > Conceptually PLN was part of the cognitive architecture design that > > led to the creation of OpenCog framework in the first place. But in > > terms of the actual software architecture of the framework PLN > doesn't > > have any distinguished role... > > > > > > > For example, if it is about natural language and resolving > > ambiguities, it could still be done by keeping a sum type in > > relevant position inside syntax tree that can be analyzed and > > decided afterwards, again using some non-low-level AtomSpace > > constructs, even possibly a PLN itself defined in a terms of > AtomSpace. > > > > PLN rules are defined in terms of Atomspace > > > > PLN truth value formulas are defined as external functions wrapped in > > GroundedSchemaNodes, but that's a temporary efficiency > optimization... > > which could also be done for any other rule-set btw... > > > > 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] > > <mailto:opencog%[email protected]>. > > To post to this group, send email to [email protected] > > <mailto:[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/CACYTDBeAQdQOOb1ytnSfcC323a3%3DiL%2BddBHyUywM2OfYZi2FjQ%40mail.gmail.com > . > > For more options, visit https://groups.google.com/d/optout. > > > > -- > > 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] > > <mailto:[email protected]>. > > To post to this group, send email to [email protected] > > <mailto:[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/CAB5%3Dj6XJ9R0tkHbvq4EvXZQcBFd4eQVkAw-GjJNEm22USM8DSA%40mail.gmail.com > > < > https://groups.google.com/d/msgid/opencog/CAB5%3Dj6XJ9R0tkHbvq4EvXZQcBFd4eQVkAw-GjJNEm22USM8DSA%40mail.gmail.com?utm_medium=email&utm_source=footer > >. > > For more options, visit https://groups.google.com/d/optout. > > -- > 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/d6f8912e-e51d-5b9e-03ba-ab59e4f2820a%40gmail.com > . > For more options, visit https://groups.google.com/d/optout. > -- 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/CAB5%3Dj6UADWWyCSuY67S7erm7WXENqCVm5mSCOM%2BK%3D1Hn8cP0Qg%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
