Re: consciousness, and AI KR

Dear all,
I concur with Dave Raggett's take on the subject, science and engineering do not deal with soft issues of linguistic interpretation typically found in religion, philosophy and psychology.
The following link provides an interesting article on how language shapes our formation of abstract concepts
Exploring the brain basis of concepts by using a new type of neural network
https://medicalxpress.com/news/2023-10-exploring-brain-basis-concepts-neural.html
The funny thing is, that these findings are nothing new, Buddhist philosophers have pointed this out in many forms. It is just only now that all of these areas of investigation are converging.Linguistic ambiguity and cognitive bias are no new subjects, but are only now becoming important in the context of creating AGI.
I propose sticking to the path described by Dave, but be mindful of what we come across in the process as long as it contributes to enhancing our formal knowledge representation modeling.

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 

    On Tuesday, October 24, 2023 at 10:40:03 PM AST, Patrick Logan <patrickdlogan@gmail.com> wrote:  
 
 "This also includes models of other agent’s beliefs and goals, i.e. a theory of mind.  Agents also benefit from a model of past, present and future, i.e. a functional episodic memory that complements encyclopaedic memory, such as birds fly and dogs bark. Episodic memory enables agents to reason about cause and effect, to understand intent, and to create and adapt plans"
"If consciousness is seen as too overloaded a term, then what word would be better for describing the subjective experience of artificial agents?  We could then discuss how that experience depends on different capabilities, e.g. episodic memory, theory of mind, behavioural norms, etc"
I quoted from two of Dave's messages. These seem to me aimed at more concrete discussions, whether or not we use the term "consciousness". I'm interested in the DARPA Explainable AI program and related efforts. This CG seems to have a similar direction. Not just explainability but also to some degree, openness / transportability / interoperability.
My background is completely in symbolic AI (expert systems) and knowledge representation (mainly semweb tech today). I work alongside a machine learning group and provide data for them. I am wondering how we can build better integrated systems. What are the different capabilities that can enable that?



On Tue, Oct 24, 2023 at 1:44 AM Dave Raggett <dsr@w3.org> wrote:

Hi Patrick,
My aim was to encourage analytic discussion on an AIKR perspective on consciousness rather than the many other potential perspectives.  One could argue about how to account for qualia from a philosophical perspective, but that is very different from consideration of how colours are handled in artificial neural networks, e.g. training a robot to count the number of red objects in a camera view. If consciousness is seen as too overloaded a term, then what word would be better for describing the subjective experience of artificial agents?  We could then discuss how that experience depends on different capabilities, e.g. episodic memory, theory of mind, behavioural norms, etc.  Is that of interest to you?

On 24 Oct 2023, at 00:33, Patrick Logan <patrickdlogan@gmail.com> wrote:
There are several terms here without even semi-formal definitions that are doing a lot of work, i.e. your claims are vague and difficult to discuss clearly let alone measure and assess.
Given the wide berth of interpretation it's especially bold to claim a false dichotomy of either one agrees with your "facts" or one is relying on "faith".
On Mon, Oct 23, 2023, 10:42 AM Dave Raggett <dsr@w3.org> wrote:

From the AI KR and computational view consciousness isn’t a hard problem.  Subjective experience distils to information processing with systems of neurons. Redness is just a vector of neural activation. Agents have situational awareness, i.e. a model of their current environment and goals, enabling them to decide on what actions to take. This also includes models of other agent’s beliefs and goals, i.e. a theory of mind.  Agents also benefit from a model of past, present and future, i.e. a functional episodic memory that complements encyclopaedic memory, such as birds fly and dogs bark. Episodic memory enables agents to reason about cause and effect, to understand intent, and to create and adapt plans.

However, this won’t convince everyone.  Plenty of people have beliefs that are a matter of faith rather than of facts. That’s fine. But engineering and science doesn’t work that way!  AI will continue to evolve and AGI is just a matter of time.  I attach a picture that makes the point. A stochastic synthesis of ideas as evidence that artistic sensibility can be reduced to neural processing.


> On 22 Oct 2023, at 05:38, Paola Di Maio <paola.dimaio@gmail.com> wrote:
> 
> 
> Consciousness is too huge a topic . Undecidable, too much can be said about without ever reaching any conclusion, possibly because no single theory or point of view can exhaust the subject. However
> I d like to suggest simply that it is tackled only in relation to AI KR. Surely. consciousness is relevant to AI and to KR discussion and potential standards. We should keep that in mind where possible and parsimoniously limit our considerations accordingly
> 
> I ll leave it to Carl to liaise with the WoT group, since he is a member there and brought up the subject.
> I ll work on tidying up some of the resources shared on the list into some form of coherent narrative when I can, that is my next task

Dave Raggett <dsr@w3.org>





Dave Raggett <dsr@w3.org>