Re: Semantic / CogAI filesystem?

I wonder what the results would be without human "fine tuning"..

Language colonialism.

This issue about the meaning of words, needs to be addressed somehow, IMO.

The more people talk about the new definition of web 3.0 the more it sounds
like a bunch of w3c peeps are Satoshi, being that they must have invented
blockchain if they were all working on it way back then...

Tim.h.

On Tue, 14 Nov 2023, 11:18 pm Ronald Reck, <rreck@rrecktek.com> wrote:

> Yes, I did the same exercise yesterday and unsurprisingly got the same
> results.
> It seems to me that the distinction lies in the role of human curation of
> results.
>
>
> On Tue, 14 Nov 2023 19:47:49 +1000, Timothy Holborn <
> timothy.holborn@gmail.com> wrote:
>
> > I've 'ChatGPT'd' (v4) the question;   "If LLMs are now defined as stuff
> > like GPT / Transformer models, What do you / w3c call older LLMs like;
> > Babelnet, cyc, FrameNet, Framester, Lemon, linguistlist, SUMO, WordNet,
> > etc..."
> >
> > Response is:
> >
> > The technologies you've mentioned, such as BabelNet, Cyc, FrameNet,
> > Framester, Lemon, Linguist List, SUMO (Suggested Upper Merged Ontology),
> > and WordNet, are generally categorized under the umbrella of Natural
> > Language Processing (NLP) tools and resources, rather than being
> classified
> > as Large Language Models (LLMs) like GPT or Transformer models. These
> older
> > systems are often referred to by more specific terms based on their
> > function and structure:
> >
> > *Lexical Databases or Ontologies:* WordNet is a prime example of a
> lexical
> > database, while SUMO is an ontology. These systems provide structured
> > information about words, concepts, and their relationships.
> >
> > *Semantic Networks:* BabelNet can be considered a semantic network as it
> > integrates lexical, conceptual, and ontological information.
> >
> > *Knowledge Bases or Knowledge Graphs:* Cyc is an example of a knowledge
> > base, which is a system that stores information in a structured form,
> often
> > used for AI and NLP applications.
> >
> > *Linguistic Frameworks*: FrameNet is a project that catalogs lexical and
> > semantic information based on frame semantics. It's more of a linguistic
> > resource than a computational model.
> >
> > *Linked Data and Ontology Frameworks:* Framester and Lemon (Lexicon Model
> > for Ontologies) are related to linked data and ontology frameworks. They
> > provide ways to integrate linguistic information with semantic web
> > technologies.
> >
> > *Linguistic Resources or Databases:* The Linguist List falls under the
> > category of linguistic resources, providing comprehensive references for
> > linguistic information.
> >
> > The distinction between these resources and LLMs like GPT or Transformer
> > models lies in their function and underlying technology. The older
> systems
> > are primarily focused on structured representation of linguistic or
> > ontological information, often relying on manually curated data or
> > rule-based approaches. In contrast, LLMs use machine learning (especially
> > deep learning) to generate or understand natural language, relying on
> > large-scale data and neural network architectures.
> >
> > In summary, while the W3C (World Wide Web Consortium) and other entities
> > involved in web standards and technologies don't have a specific
> collective
> > term for these older systems, they are typically referred to based on
> their
> > specific function (e.g., lexical databases, ontologies, knowledge bases)
> in
> > the broader field of NLP and AI.
> >
> >
> ___________________________________________________________________________________________________________________________________
> >
> >
> > So, if i want to make a clear distinction between an LLM (that requires
> an
> > enormous amount of compute power to operate 'privately'); and the means
> to
> > create AI resources, by updating stuff like BabelNet, Cyc, FrameNet,
> > Framester, Lemon, Linguist List, SUMO (Suggested Upper Merged Ontology),
> > and WordNet to support spatio-temporal 'ai datasets' that can be used
> > locally (offline, although offline does have impacts / implications) with
> > limited hardware resources; then, what should that field of work be
> called?
> >
> > Perhaps 'knowledge base(s)'??
> >
> > I've started a new document for 'definitions':
> >
> https://docs.google.com/document/d/1ZEBQ4zs-yb4idpIqR06vvGhOX83G0P6LK3YeD12NwPU/edit?usp=sharing
> >
> >
> > It would be great if we could work on defining a range of terms that have
> > express meanings defined via URIs, etc. That could be shared across
> > multiple groups / venues, etc.
> >
> > I expect some should be easier than others...  Web 3.0 & "human centric
> ai'
> > perhaps being amongst the more difficult ones to define atm, imho..
> >
> > Tim.H.
> >
> >
> > On Tue, 14 Nov 2023 at 18:35, Timothy Holborn <timothy.holborn@gmail.com
> >
> > wrote:
> >
> > >
> > >
> > > FWIW: I've started on:
> > >
> > >
> > >
> https://docs.google.com/document/d/17VHMpoIXbBT6AXkspPDkHfoTi80GL6OwiAPs9VrHVbY/edit?usp=drivesdk
> > >
> > > But it'll be updated ALOT over the coming days.  It would be good if
> > > others also put together their thoughts, then perhaps we could
> socialise
> > > considerations..  if wanted.
> > >
> > > NB also: I really like your slides, the presentation quality is great.
> > >
> > > Best.
> > >
> > > Tim.h.
> > >
> > >
> > >
> > >
> > > On Tue, 14 Nov 2023, 6:28 pm Dave Raggett, <dsr@w3.org> wrote:
> > >
> > >> Naming is often referred to as bike shedding …
> > >>
> > >
> > > https://en.wikipedia.org/wiki/Law_of_triviality
> > >
> > > https://wikidata.org/wiki/Q169899
> > >
> > >
> > >> We need names, but it often takes up a lot of time better spent on
> other
> > >> things.
> > >>
> > >
> > > Yeah, does take alot of time..
> > >
> > >
> > >
> > >> On 13 Nov 2023, at 19:27, Timothy Holborn <timothy.holborn@gmail.com>
> > >> wrote:
> > >>
> > >> quick questions;
> > >>
> > >> If LLMs are now defined as stuff like GPT / Transformer models,
> > >>
> > >> What do you / w3c call older LLMs like; Babelnet, cyc, FrameNet,
> > >> Framester, Lemon, linguistlist, SUMO, WordNet, etc...
> > >>
> > >> Also, is web 3.0 still web 3.0?  it seems to have been adopted in some
> > >> way, which kinda leaves open the opportunity to define internet 3.0 ;)
> > >>
> > >> whilst the 'web 3.0' stuff isn't so important atm, the issue about
> 'large
> > >> language models;' vs. transformer models, or however the terminology
> should
> > >> now be categorised; which thereby also leads to various issues about
> how
> > >> the term 'ai' is defined, but starting small....
> > >>
> > >> https://www.w3.org/TR/wordnet-rdf/  and all such things seemed to be
> > >> fairly big jobs.  Whilst I think they need to be updated to support
> > >> spatio-temporal notations, vector stuff, etc... (resources for logical
> > >> programming, Computational geometry, et.al. imho, ) It seems there's
> a
> > >> problem with being clear about terms and their meanings...
> > >>
> > >> FWIW: Figure it's probably an important issue to raise in relation to
> W3C
> > >> activities generally.
> > >>
> > >> cheers,
> > >>
> > >> tim.h.
> > >>
> > >>
> > >> On Mon, 13 Nov 2023 at 20:17, Timothy Holborn <
> timothy.holborn@gmail.com>
> > >> wrote:
> > >>
> > >>> K.
> > >>>
> > >>> I'll follow-up when I've put together some thoughts, following more
> > >>> research.
> > >>>
> > >>> Cheers.
> > >>>
> > >>> Tim.h.
> > >>>
> > >>> On Mon, 13 Nov 2023, 8:11 pm Dave Raggett, <dsr@w3.org> wrote:
> > >>>
> > >>>>
> > >>>>
> > >>>> On 13 Nov 2023, at 10:37, Timothy Holborn <
> timothy.holborn@gmail.com>
> > >>>> wrote:
> > >>>>
> > >>>> Hi Dave,
> > >>>>
> > >>>> I wondered if you had any thoughts on the potential application of
> your
> > >>>> CogAI work to create a semantic/CogAI filesystem?
> > >>>>
> > >>>>
> > >>>> Yes, and in conjunction with large language models.  I see the
> > >>>> combination (something I refer to as cognitive databases) as an
> > >>>> evolutionary replacement for today’s graph databases, and paving
> the way to
> > >>>> zero-code applications. I see this as enabling collaborative
> knowledge
> > >>>> engineering (see my talk [1]) where the computer and human play
> > >>>> complementary roles.  The human can ask the computer to perform an
> analysis
> > >>>> of the data, leaving the details to the computer to figure out for
> itself.
> > >>>>
> > >>>> The huge hype around generative AI has drawn attention away from the
> > >>>> limitations of the current approaches to large language models.  I
> see many
> > >>>> opportunities for evolving artificial neural networks inspired by
> what we
> > >>>> know about human cognition. This will enable computers to become
> very
> > >>>> effective partners for human-machine collaborative work.
> > >>>>
> > >>>> Neural networks and vector spaces are very powerful for representing
> > >>>> complex statistical relationships, as a generalisation of symbolic
> > >>>> knowledge, and learned and queried by machine. This changes the
> conception
> > >>>> of databases considerably. The impact of this will be dramatic.
> > >>>>
> > >>>> [1] http://www.w3.org/2023/10/10-Raggett-AI.pdf
> > >>>>
> > >>>>
> > >>>> I was looking at the solid code again recently, and was thinking
> about
> > >>>> the implementation structure.
> > >>>>
> > >>>> I haven't done much more thinking about it yet, but thought I'd ask
> you
> > >>>> just in case.
> > >>>>
> > >>>> Whilst my plans for human centric AI stuff, particularly my
> thoughts on
> > >>>> my own implementation, may not be solid, although I want
> compatibility /
> > >>>> backwards compatibility...
> > >>>>
> > >>>> It seemed to me that there's an ability to make distinctions between
> > >>>> different agents using multiple "pods" via methods that assume
> domain
> > >>>> ownership.
> > >>>>
> > >>>> Without getting into it further, the basic thought was about
> semantic
> > >>>> file systems which then led to wondering about CogAI file
> systems...  How
> > >>>> that might work?
> > >>>>
> > >>>> Tim.H.
> > >>>>
> > >>>>
> > >>>>
> > >>>>
> > >>>> Dave Raggett <dsr@w3.org>
> > >>>>
> > >>>>
> > >>>>
> > >>>>
> > >> Dave Raggett <dsr@w3.org>
> > >>
> > >>
> > >>
> > >>
>
>
> Ronald P. Reck
>
> http://www.rrecktek.com - http://www.ronaldreck.com
>