Re: W3 AIKR CG

Peter, and all thanks for chiming in
The AI KR CG has had many contributions from stratml proponents because of
Owen's has been consistent in posting and promoting his standard via this
list. Of course the mission of AI KR is far greater, yet none of us can
make a commitment to a single agenda. So far we have tried to keep up to
date with matters that impact AI KR so that when the time comes, we can be
at a minimum up to scratch,
So Peter, thanks for sharing the ideas and it would be good if these could
be rounded off as action points.
I apologise I have been silent, I have been too busy and short of resources
I promised that I would share my latest work after publication, and this is
often in the hands of the publisher for a long time. We do have an ISBN and
publication is imminent, so let me anticipate that the proposal I have
consists of......... using a metamodel approach/metamodel card to capture
AI KR construct to gude system development and implementation. I ll share
the chapter soon.
On another front, I am researching AI in Art Curation.
I have a couple of weeks beginning May during which I expect to have some
spare
cycles and may ping an iMeet invite to the list, and will put together a
few slides for each with pointers.
What can this AI KR CG can do after that, it is up to its members.
I may be unable to follow matters much beyond this point until further
notice as I need to put my energy in personal and family matters, but
will be glad to see members pick up on what is left doing
PDM


On Thu, Apr 13, 2023 at 5:28 PM Peter Rivett <
pete.rivett@federatedknowledge.com> wrote:

> I agree that the AI KR Group could use some re-booting but not IMO as the
> "StratML for AI" Group.
> I'm interested in the representation of knowledge more broadly than plans.
> And I thought that was the intent of the Group when I joined.
>
> Specifically for LLMs and generative AI, I'm interested in knowledge
> representation topics such as:
>
>    - how to represent the knowledge represented in training data
>    (knowledge about the knowledge). So that the AI trained has full context of
>    the provenance and scope of that data and hence its applicability to
>    solving different problems. For example to know that a set of data captured
>    from US college students may not be applicable to elders in Japan.
>    - also any permission information associated with that data. For
>    example both artists and programmers may not want to see their work
>    repurposed for no recompense or credit to them
>    - how to set and document rules governing what the AI may not do (e.g.
>    generating racist or pornographic content)
>    - how the provenance can be represented in a standard way to support
>    references back from any solutions generated by the AI, and potentially
>    support the legal auditing that may increasingly be required
>    - how to represent and advertise the availability of *current*
>    knowledge that may be used by the AI to supplement the training data with
>    more recent information (e.g. current ChatGPT training data is from 2021)
>    - and *fact check* solutions to eliminate the hallucinations
>    - (more future looking) how to combine knowledge graphs with machine
>    learning e.g. through Graph Neural Networks, vector embeddings and the like
>
> That said, I don't have a lot of time in my schedule so would see myself
> more as a reviewer, integrator, and consumer than an active contributor of
> content.
> Though count me out of anything focused only on StratML. Sorry, that just
> does not interest me.
>
> Pete
>
> PS trying to represent strategy/plans at the level of generic software
> like ChatGPT is already old hat. People are now using tech like Auto-GPT
> https://www.digitaltrends.com/computing/what-is-auto-gpt/ to set goals
> for their own single-purpose agents, including via dialog with the user.
>
> PPS if you want a list of companies doing something with AI try the 21k
> here from my own company's KG
> https://golden.com/query/list-of-artificial-intelligence-companies-MVG
>
>
> Pete Rivett
> Ontologist, Golden
> Director, Enterprise Knowledge Graph Foundation
> Schedule a meeting at https://calendly.com/rivettp
>
>
>
> ------------------------------
> *From:* Owen Ambur <owen.ambur@verizon.net>
> *Sent:* Wednesday, April 12, 2023 7:44 PM
> *To:* ProjectParadigm-ICT-Program <metadataportals@yahoo.com>; W3C AIKR
> CG <public-aikr@w3.org>; carl mattocks <carlmattocks@gmail.com>
> *Cc:* Naval Sarda <nsarda@epicomm.net>; pradeep.jain@ictect.com <
> pradeep.jain@ictect.com>; William Glascoe III <eosocxo@comcast.net>;
> chris@chriscfox.com <chris@chriscfox.com>; Jorge Sanchez <jorge@vionta.net
> >
> *Subject:* Re: W3 AIKR CG
>
> Milton & Carl, I'll be happy to participate in any effort to leverage the
> StratML standard to help make sense of what is taking place with ML/AI, in
> order to help ensure not only that humans are "in the loop" but also that
> our interests are being well served.
>
> From my perspective, perhaps the best thing we might try to accomplish is:
>
> a) encourage the developers of such applications to document their own
> strategic plans and performance reports in StratML format or at least to
> share the information required for such plans in some format;
> b) if they fail to do so, to do our best to do it for them, leveraging the
> listings of such applications that have been compiled by others; and
> c) make such plans available for discovery in a service like
> https://search.aboutthem.info/
>
>
> Here are some references that might provide a basis from which to begin:
>
>
> https://www.simplilearn.com/tutorials/artificial-intelligence-tutorial/artificial-intelligence-applications
>
> https://www.projectpro.io/article/10-awesome-machine-learning-applications-of-today/364
>
> https://medium.com/app-affairs/9-applications-of-machine-learning-from-day-to-day-life-112a47a429d0
> https://idapgroup.com/blog/top-machine-learning-applications/
> https://intellipaat.com/blog/top-artificial-intelligence-tools/
> https://www.wordstream.com/blog/ws/2017/07/28/machine-learning-applications
> https://research.aimultiple.com/ai-usecases/
> https://builtin.com/artificial-intelligence/ai-companies-roundup
>
>
> The latter lists 68 AI companies.
>
> If we were to take up such an effort, it might be good to engage the AI
> applications themselves in helping us to identify their missions, visions,
> values, goals, objectives, and stakeholders as well as the plans of others,
> to the degree they may be aware of each other.
>
> See, for example, the plans that ChatGPT and You.com revealed upon
> prompting, at https://stratml.us/docs/CGPT.xml &
> https://stratml.us/docs/YOU.xml
>
> See also this exchange with You.com:
>
> *Can you outline the goals and objectives of ChatGPT?*
>
> Yes, ChatGPT is a natural language generation tool that can be used for a
> variety of purposes. Some potential goals and objectives of ChatGPT could
> include:
>
> Improving productivity by automating routine tasks and generating reports
> or other documents
>
> Providing assistance with writing, editing, or proofreading content
>
> Generating engaging and informative content for social media, blogs, or
> websites
>
> Supporting customer service by providing quick and accurate responses to
> common queries
>
> Streamlining communication between team members by automatically
> summarizing meeting minutes or generating progress reports Overall, the
> main goal of ChatGPT is to leverage the power of large language models to
> provide valuable assistance and improve efficiency in a variety of
> contexts.
>
> *How do they differ from your own?*
>
> As an AI language model and not a human, I do not have personal opinions
> or goals. However, I can tell you that the goals and objectives of ChatGPT
> may differ from those of individual users or organizations based on their
> specific needs and contexts. ChatGPT is a tool that can be used for a wide
> range of purposes and can be customized to fit the specific requirements of
> different use cases. Some users may want to use ChatGPT for generating
> content or automating routine tasks, while others may want to use it for
> customer service or communication purposes. Ultimately, the success of
> using ChatGPT will depend on how well it aligns with the goals and
> objectives of the user or organization.
>
> *You previously revealed your goals and objectives as outlined at
> https://stratml.us/docs/YOU.xml <https://stratml.us/docs/YOU.xml>, but can
> you explain how your value proposition differs from ChatGPT?*
>
> I apologize for the confusion but as an AI language model, I do not have
> personal goals or a value proposition. My purpose is to provide assistance
> and answer questions to the best of my ability based on the input provided
> to me. ChatGPT is a specific implementation of a language model, and its
> value proposition may differ based on its specific capabilities and the
> goals and objectives of its creators and users. However, both myself and
> ChatGPT share the overall goal of providing valuable assistance and
> improving efficiency in a variety of contexts through the power of natural
> language generation.
>
>
> Here's a brief exchange with ChatGPT:
>
> *Can you explain the goals and objectives of You.com and how they relate
> you your own?*
>
> You.com is a search engine and personal data platform founded by Jaron
> Lanier with the goal of giving individuals control over their personal data
> and digital identity. The objectives of You.com are to provide users with a
> more personalized and secure online experience while also allowing them to
> monetize their data if they choose to do so.
>
> As an AI language model, my primary goal is to provide helpful and
> accurate responses to the questions and prompts given to me by users. I do
> not collect or store personal data, as my purpose is purely to assist with
> information and language-related tasks.
>
> While You.com and I both aim to provide users with personalized
> experiences, our methods and objectives are quite different. You.com
> focuses on giving individuals control over their personal data, while I am
> designed to provide general language processing and information retrieval
> services.
>
> To the degree that one may be providing incomplete or misleading
> information about the other, that might create incentive for them to do a
> better job of sharing their own plans in StratML format.  Creating such
> incentive is part of my intent for the AboutThem.info
> <https://aboutthem.info/> domain.
>
> If the W3C were to officially take up such an effort, it seems like the
> result could be quite positive and perhaps transformative.
>
> Owen Ambur
> https://www.linkedin.com/in/owenambur/
>
>
> On Wednesday, April 12, 2023 at 08:49:00 AM EDT, carl mattocks <
> carlmattocks@gmail.com> wrote:
>
>
> Milton
>
> I would be willing to lead some meetings dedicated to how StratML planning
> of  KR that would determine the role of Human -in-the Loop for  (1) Generative
> AI  (2)  Language Models (3) Algorithms (4) etc
>
> Carl Mattocks
>
> It was a pleasure to clarify
>
>
> On Tue, Apr 11, 2023 at 5:28 PM ProjectParadigm-ICT-Program <
> metadataportals@yahoo.com> wrote:
>
> Dear Paola, Carl and Owen,
>
> I have been posting about the current issues with AI, in particular
> generative LLMs.
>
> I think it would be instructive and useful to focus on producing a
> generalized outline for AI KR.
>
> I myself have been very busy at my end setting up a national science
> research foundation in Aruba, and am currently in the process of
> streamlining all other NGO activities I am involved with to focus solely on
> applied research.
>
> The W3 AIKR Community Group needs some rebooting, and after the skirmishes
> with Dan Brinkley I noticed that some 50 plus posts were deleted for the
> chat, because one day there were 90 posts and the next it was reduced to 35
> something with an entire timeline erased.
>
> I was wondering if we could get back to the original focus and use StratML
> to outline timeline, efforts and deliverables?
>
> Milton Ponson
> GSM: +297 747 8280
> PO Box 1154, Oranjestad
> Aruba, Dutch Caribbean
> Project Paradigm: Bringing the ICT tools for sustainable development to
> all stakeholders worldwide through collaborative research on applied
> mathematics, advanced modeling, software and standards development
>
>