Rob Yes, I understand the difference between a video and a paper. I did not think John criticized it either. If he had, it's helathy to render critique. You referenced the paper, and it remains relevant to the topic, regardless of the medium accessed.
Indeed, I was referring to your decoupling point Rob. As you are well aware, technically, there's a significant different between a relation and an association. Relation implies functional dependency, whereas the paper seems to indicate a method for relating an object to itself, and purely by abstracted value, comparing it to any other object within its universe. There are a significant number of inherent patterns right there. I value your point on position-based categories. However, if these positions are accurately plotted on a spacetime continuum, then we're looking at a very-interesting approach to enabling category theory (and like you I'm still learning) with accepted, quantum referencing within specific, topological structures. Some of us think that's exactly how the known universe achieves equilibrium. No, I'm not equating 2nd order predicate calculus, or logic to category theory. My point on "order" relates to the notion of emerging such from the calculated values of objects (points) relative to other points, within x space. This possibly may well offer a useful workaround for getting stuck in Heisenberg's Uncertainty. Once the positions of data-bearing objects are known, in a sense it would negate the need to use globalization and rather move to specifics. In my view, this would provide an environment in which deabstraction and optimization would functionally fit into. Again, I understand your preference for clarity on relationships, but relationships change. My view would be more to get a valid and reliable fix on compound relationships (in the sense of associations), where all possible changes between objects are accommodated within hierarchies of control. I think the paper made that point, probably using other terms. A change in a historical relationship, wouldn't necessarily have to mean destroying all associated historical predictions (statistically calculate). The approach in the paper seemingly allows for rapid reintegration without loss of any data. I see pure data objects. Probably, because the robustness of the system wouldn't get compromised if a point pseudo-randomly moved from one position to another. Agreed, I also like category theory, but probably for slightly different reasons to you. I'm biased towards contexts. There's more to "language" as I put it, than grammar. The paper did not mention it as such. Last, you asked: "How do you relate their relational encoding to regression?" This is an excellent question. I don't quite know how they do their relational encoding. As for regression, if I understand it correctly, it relies on functional dependencies to emerge most-probable results, as implied statistically. What I think is that, with their way of setting up each object as its own part/entity/element, they could probably relate objects statistically by the degree of overall and core similarity. If so, this would enable a fractal-relational principle for data cohesiveness. I believe cohesiveness has again become all the rage. To conclude, my excitement at what the paper contains is not for, or against any theory. Theories are great. We read and think about them. Heck, I even have a theory or two. Even so, theory must be tempered by practical results. I'd put the results in the paper as having significant empirical value. And exactly fo that reason, I could find practical value in it beyond what was stated. Why not discuss it even further? I concede that the conversation among you are quite theoretical. Even so, we see what our eyes see. I see the paradigm shift in specifying a practical, data approach to systematically converge all of engineering towards quantum fundamentals. I remain an advocate for quantum engineering methodologies and practices. It not only gives me hope that the road to purely-machine-based AGI applications could be shortened significantly. It also give me hope for commercialized products, such as "Engineering on a chip". On Thu, May 23, 2024 at 7:27 AM Rob Freeman <chaotic.langu...@gmail.com> wrote: > On Thu, May 23, 2024 at 10:10 AM Quan Tesla <quantes...@gmail.com> wrote: > > > > The paper is specific to a novel and quantitative approach and method > for association in general and specifically. > > John was talking about the presentation James linked, not the paper, > Quan. He may be right that in that presentation they use morphisms etc > to map learned knowledge from one domain to another. > > He's not criticising the paper though. Only the presentation. And the > two were discussing different techniques. John isn't criticising the > Granger et al. "relational encoding" paper at all. > > > The persistence that pattern should be somehow decoupled doesn't make > much sense to me. Information itself is as a result of pattern. Pattern is > everything. Light itself is a pattern, so are the four forces. Ergo. I > suppose, it depends on how you view it. > > If you're questioning my point, it is that definition in terms of > relations means the pattern can vary. It's like the gap filler example > in the paper: > > "If John kissed Mary, Bill kissed Mary, and Hal kissed Mary, etc., > then a novel category ¢X can be abduced such that ¢X kissed Mary. > Importantly, the new entity ¢X is not a category based on the features > of the members of the category, let alone the similarity of such > features. I.e., it is not a statistical cluster in any usual sense. > Rather, it is a “position-based category,” signifying entities that > stand in a fixed relation with other entities. John, Bill, Hal may not > resemble each other in any way, other than being entities that all > kissed Mary. Position based categories (PBCs) thus fundamentally > differ from “isa” categories, which can be similarity-based (in > unsupervised systems) or outcome-based (in supervised systems)." > > If you define your category on the basis of kissing Mary, then who's > to say that you might not find other people who have kissed Mary, and > change your category from moment to moment. As you discovered clusters > of former lovers by fits and starts, the actual pattern of your > "category" might change dramatically. But it would still be defined by > its defining relation of having kissed Mary. > > That might also talk to the "regression" distinction. Or > characterizing the system, or indeed all cognition, as "learning" > period. It elides both "similarity-based" unsupervised, and > supervised, "learning". The category can in fact grow as you "learn" > of new lovers. A process which I also have difficulty equating with > regression. ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/T682a307a763c1ced-M45b4c5c739faf7c61b229575 Delivery options: https://agi.topicbox.com/groups/agi/subscription