Here is an example of a generic ontology, which is quantum ready and supports 1-step mutation.
It could be "flavoured" for AGI purposes via content. From: a...@listbox.com To: a...@listbox.com Subject: RE: [agi] Knowm - Machine Learning Coprocessor Date: Sun, 15 Mar 2015 22:31:38 +0200 Seems there is a semantic difference here with Cyc's OWL. They do have an ontology, but what they publish is more of an ontological taxonomy, or the classification sub-component of a general, ontological component referring to Convention. This implies that somewhere, there must be a whole ontology, which unfortunately they have elected not to share. Given this example, there is no way of knowing how the ontology would deal with single-step mutation, if at all. Therefore, in its absence, it cannot be considered as a candidate AGI ontology. For AGI, below is more of a systems example of what I am talking about. Ontology is the philosophical study of the nature of being, becoming, existence, or reality, as well as the basic categories of being and their relations. Traditionally listed as a part of the major branch of philosophy known as metaphysics, ontology deals with questions concerning what entities exist or can be said to exist, and how such entities can be grouped, related within a hierarchy, and subdivided according to similarities and differences From: a...@listbox.com To: a...@listbox.com Subject: RE: [agi] Knowm - Machine Learning Coprocessor Date: Sun, 15 Mar 2015 12:28:43 -0700 Try this... http://www.cycfoundation.org/ There are links for the concept browser, etc. I think you can download the ontology here: http://datahub.io/dataset/opencyc You'll have to figure out how it handles mutation. Cheers, Michael. From: a...@listbox.com To: a...@listbox.com Subject: RE: [agi] Knowm - Machine Learning Coprocessor Date: Sun, 15 Mar 2015 21:00:02 +0200 PM Thanks for the headsup. I'd appreciate a link to a logical system's model of the whole ontology. How does it handle mutation (in the narrowest sense) as an adaptive construct? From: a...@listbox.com To: a...@listbox.com Subject: RE: [agi] Knowm - Machine Learning Coprocessor Date: Sun, 15 Mar 2015 11:03:14 -0700 Why not use Cyc's ontology then? It's been over 30 years in the making. ~PM From: a...@listbox.com To: a...@listbox.com Subject: RE: [agi] Knowm - Machine Learning Coprocessor Date: Sun, 15 Mar 2015 19:10:51 +0200 Further to the previous comment on ontology. Herewith, my thoughts... One simply cannot put the cart before the horse, not if it is expected of the horse to pull the cart. The above does somehow return us to more than just the ontological and complexity questions. It semeingly returns us to a specific vision of AGI, reflected as a paradigm of reasoning and associated mindset, from which should flow a natural AGI ontology. Suppose such an AGI ontology was viewed as a system, in compliance with systems theory. From this would follow that, unless it does not specify the approach to effectively address definite AGI research objectives with, it would ultimately fail. Doe the previous statement imply there might be more than one ontology that would effectively relate to AGI? In this context then, yes and no. No, on condition that the AGI ontology was specified at an appropriate AGI level (the AGI bar everyone is talking about). Yes, on condition that the AGI bar was set high enough to embrace emerging and existing ontological value within its systems boundary. For AGI, I contend this point to be of critical value. As such then, if an AGI ontology was designed at the highest-possible AGI level, then its followers may rely on it to become the holistic containerform for whatever form, function, association, entanglement, context or content may be invoked within its system boundaries. In short, it would prove to be valid and reliable, as a scientific basis for governing AGI with. Therefore, one of the starting points of any ontology should be the philosophical sub-model or collective consciousness that is best suited for a future construct of AGI. For example, we may start by returning to the fathers of AI and gather their thoughts on the philosophic approach, and so on. From such an eventual collection of entangled thoughts, as a philisophical foundation model, would emerge the way (in the sense of a path of enlightenment) towards an adaptive AGI model. I'm adding the term "adaptive", because we must surely entangle the philosophy of adaptiveness within a standard model of AGI. This has to be done from the outset, first to prevent followers of the path to lose their AGI way. Further, to inform all analysis-and-design and development decisions along the way, of every insantiation of the AGI way, in any capacity of a selected AGI SDLC. In the case of such a term as "adaptive", an AGI ontology would be further developed then to releft adaptiveness, as a meta model of adaptiveness, as inherent, economic value. Further, once the ontology is in place, the principle of adaptivenesswould naturally become a meta standard - or meme - for an emerging culture of AGI. This then, within the notion of an AGI ontology itself, but not conclusively so, merely presenting as a possible scenario. Assuming then such an ontologic was completed as a standard AGI model, when applying the ontology to a specific AGI problem of adaptiveness (by following the AGI way), this model would systemically invoke the principle of adaptivenes across all its contextual components, and in doing so, satisfy the inherent meta requirements to emerge the value as a possible solution to the problem in a form of functional (reductionist) adaptiveness for AGI. I think, many of the circular debates around coding and languages, although most useful in their own right, intelligent and relevant for functional application of adaptiveness, would unfortunately all meet their respective cul de sac. I'm saying this purely as an argument in favour of the definition and adoption of an appropriately specified AGI ontology, as a way of AGI. From: a...@listbox.com Date: Sun, 15 Mar 2015 16:19:51 +0100 Subject: Re: [agi] Knowm - Machine Learning Coprocessor To: a...@listbox.com On Sat, Mar 14, 2015 at 10:02 PM, Steve Richfield via AGI <a...@listbox.com> wrote: IMHO the future of ALL AGI approaches lies in the careful design of the APIs and other interfaces As much as I like to dabble in all the freely available designs and codebases, and with the obvious "commoditization" that widely available APIs will bring, I dont see how this would solve the issues of ontological and complexity barriers of AGI. More specifically, in my own designs maintaining a well grounded ontology at all times is a sine qua non, as are recursive agent interactions ("game trees" if you will), with agents of course well grounded and maintained across the different scenarios (inevitable combinatorial explosion with probabilistic agents). No, I do not expect any API magic! But yeah, since no "learning" problem could be assummed to be of one type or another, a generic hypothesis engine or, in should that prove too tricky, a large collection of machine learning APIs and automation is also a vital AGI component, AT AGI | Archives | Modify Your Subscription AGI | Archives | Modify Your Subscription AGI | Archives | Modify Your Subscription AGI | Archives | Modify Your Subscription AGI | Archives | Modify Your Subscription AGI | Archives | Modify Your Subscription ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
STEP10.wmf
Description: windows/metafile
