On Fri, Sep 2, 2016 at 11:19 PM, Ben Goertzel <[email protected]> wrote:
> Hmm, I don't feel like I'm brushing you off. I'm actually trying to > understand why you think linear or affine logic is needed --- I don't > get why, I suspect you have some intuition or knowledge here that I'm > not grokking, and I'd honestly love more > clarification/elaboration/explanation from you... > Well, the blog post http://blog.opencog.org/2016/08/31/many-worlds-reasoning-about-reasoning/ really tries to lay down the most basic ideas. As you read it, I believe that you should be reminded of the old idea of "inference trails" which I diligently tried not to mention, because I am afraid it will confuse the issue. If you read that, and think "how could one implement multiple atomspaces efficiently", then, yes, ContextLink and/or CompositeTruthValue is a way of doing that. I claim that all the high-falutin mathematical terminology and concepts apply to this particular situation, but discussing that further, just right now, seems unproductive. So let me set that aside. But I do have the general sense that, as we move data through the system, using either ContextLinks, or something else that implements the multiverse model, would be useful. There is a natural, obvious point of contact for this: every parse given to you, by relex, is effectively its own "universe", already. That's just the way LG and relex were designed. Its how just about all parsers are designed. The classic example is "I saw the man with the telescope", which has two parses, each parse is its own universe. One parse is: _subj(see, I) _obj(see, man) _advmod(see, with) _pobj(with, telescope) the other parse is nearly identical: _subj(see, I) _obj(see, man) _prepadj(man, with) _pobj(with, telescope) We have several choices with how to deal with this. We can try to keep these in separate "universes", maybe using ContextLinks (for example -- there are other ways, too), and then accumulate evidence until we can rule out one of these interpretations. The other approach, that Nil was advocating with his distributional-TV proposals, is to jam these two into one, and say that _advmod(see, with) is half-true, and _prepadj(man, with) is half-true, -- and then somehow hope that PLN is able to eventually sort it out. We currently don't do this approach, because it would break R2L -- the R2L rules would probably mis-behave, because they don't know how to propagate half-truths. My gut instinct is that keeping the two different interpretations of "I saw the man with the telescope" separate, for as long as possible, is better. Using the distributional-TV to (prematurely) merge them into one will probably lead to confusion. However, this idea of keeping contexts separate, and when one should or should not merge them together, requires some sort of common shared vocabulary, so that we can talk about it. The existing code talks about "parses" and "interpretations" and "word instances" and, maybe that is all that is needed, for now. But I suspect that it won't be enough, not for long, because PLN inferences done on one "interpretation" almost necessarily have to be distinct from the inferences one on the other interpretations, and we don't have good control over this. Maybe ContextLinks or maybe ContextualTruthValues would be good for this. Maybe inference trails might help. I dunno. The reason I dunno is because this level of abstraction is too low; it is too hard to think and plan and design at that level. More abstraction makes understanding easier. My hope was that, by talking about these other kinds of logics, it would clarify the issues -- for example, by understanding that "parsing" and "reasoning" are really the same kind of thing, and that therefore, the various tricks and techniques and algorithms developed for parsing could be applied to reasoning as well. Backward and forward chaining are very crude, very primitive tools. Far superior algorithms have been invented. I'm quite sure that we can do a lot lot better than merely backward/forward chaining in PLN. But we can't get there until we start talking at the correct level of abstraction. -- Linas > About ContextLink / CompositeTruthValue -- an interesting relevant > question is whether we want/need to use it in the PLN backward chainer > which Nil is now re-implementing.... Quite possibly we do... > > > > On Sat, Sep 3, 2016 at 12:11 PM, Linas Vepstas <[email protected]> > wrote: > > Yes. I am starting to get very annoyed. Whenever I talk about > > CompositeTruthValue, which I did earlier, I get the big brushoff. Now, > when > > I finally was able to sneak it back into the conversation, I once again > get > > the big brushoff. > > > > I am starting to get really angry about this. I am spending wayyy too > much > > time writing these emails, and all I get is blank stares and the > occasional > > snide remark back. This is just not that complicated, but as long as > you do > > not bother to apply your considerable brainpower to all of this, the > > conversation is utterly completely stalled. > > > > I'm pretty angry right now. > > > > --linas > > > > > > On Fri, Sep 2, 2016 at 10:44 PM, Ben Goertzel <[email protected]> wrote: > >> > >> Linas, > >> > >> On Sat, Sep 3, 2016 at 10:50 AM, Linas Vepstas <[email protected]> > >> wrote: > >> > Today, by default, with the way the chainers are designed, the various > >> > different atomspaces are *always* merged back together again (into one > >> > single, global atomspace), and you are inventing things like > >> > "distributional > >> > TV" to control how that merge is done. > >> > > >> > I am trying to point out that there is another possibility: one could, > >> > if > >> > desired, maintain many distinct atomspaces, and only sometimes merge > >> > them. > >> > So, for just a moment, just pretend you actually did want to do that. > >> > How > >> > could it actually be done? Because doing it in the "naive" way is not > >> > practical. Well, there are several ways of doing this more > efficiently. > >> > One way is to create a new TV, which stores the pairs (atomspace-id, > >> > simple-TV) Then, if you wanted to merge two of these "abstract" > >> > atomspaces > >> > into one, you could just *erase* the atomspace-id. Just as easy as > that > >> > -- > >> > erase some info. You could even take two different (atomspace-id, > >> > simple-TV) > >> > pairs and mash them into one distributional TV. > >> > >> I note that we used to have something essentially equivalent to this, > >> for basically this same reason..... > >> > >> It was called CompositeTruthValue, and was a truth value object that > >> contained mutliple truth values, indexed by a certain ID. The ID > >> was a version-ID not an atomspace-ID, but same difference... > >> > >> A dude named Linas Vepstas got rid of this mechanism, because he > >> (probably correctly) felt it was a poor software design ;) > >> > >> The replacement methodology is to use EmbeddedTruthValueLink and > >> ContextAnchorNode , as in the example > >> > >> Evaluation > >> PredicateNode "thinks" > >> ConceptNode "Bob" > >> ContextAnchorNode "123" > >> > >> EmbeddedTruthValueLink <0> > >> ContextAnchorNode "123" > >> Inheritance Ben sane > >> > >> which uses more memory but does not complicate the core code so much... > >> > >> -- Ben > >> > >> > >> > >> > >> -- > >> Ben Goertzel, PhD > >> http://goertzel.org > >> > >> Super-benevolent super-intelligence is the thought the Global Brain is > >> currently struggling to form... > >> > >> -- > >> You received this message because you are subscribed to the Google > Groups > >> "link-grammar" 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/link-grammar. > >> For more options, visit https://groups.google.com/d/optout. > > > > > > -- > > You received this message because you are subscribed to the Google Groups > > "link-grammar" 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/link-grammar. > > > > For more options, visit https://groups.google.com/d/optout. > > > > -- > Ben Goertzel, PhD > http://goertzel.org > > Super-benevolent super-intelligence is the thought the Global Brain is > currently struggling to form... > > -- > 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/CACYTDBcLcdnVyZupxbrQKZ%2Bikd-C49NM1DYtzWj_ojcFOBxLnw% > 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]. 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/CAHrUA35CTBBteschh0vi%3DG5je9L%2BV0UF0%3DaRkBcoTicWxMSnpA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
