Re: reader, On connectionism, Smolenski 1988

I must agree with Paola, an excellent article.
A quick glance at some of the concepts mentioned. in particular P-knowledge and S-knowledge underscore a fundamental issue in current neuro-cognitive research, as reflected in a growing body of literature on how single brain cells function, store information, and initiate or mediate in neural network processes.
It s obvious that brain cells do not store information in bits and bytes using ASCII or UniCode characters, which makes the neuro-symbolic or BICA approaches to knowledge representation for AI very hard to achieve.
I find the following article very instructive:https://www.researchgate.net/publication/358488001_Neural_Sheaf_Diffusion_A_Topological_Perspective_on_Heterophily_and_Oversmoothing_in_GNNs
Cellular sheaves and diffusion provide a mathematical and symbol based way to deal with subsymbol processing, although not complete.
What is evident from current neuroscience research is that many cognitive processes require many types of brain cells in concert to initiate, combine and filter through processes in many parts of the brain.
The mathematics required to deal with this involve algebraic topology, category theory, a wide range of AI concepts, graphs or neural nets based, and quantum physics.
When we allow for quantum physics to be part of the formal description for KR for AI, we introduce the uncertainty principle into information processing leading to a "information uncertainty principle" meaning we may never know for certain the location of information and the content of information at the same time.

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 Monday, May 30, 2022, 12:15:25 AM AST, Paola Di Maio <paola.dimaio@gmail.com> wrote:  
 
 As neuro symbolic integration is addressed in the context of KRthere are many (false ) assumptions that come in the way and some historical misunderstandings as to the definitions of symbolism, connectionism, sub symbolism 
understanding which is central to neuro symbolic integration research today

Here some essential background, to quality discussion 
Hats off to Smolenski 

BEHAVIORAL AND BRAIN SCIENCES (1988) 11, 1-74
On the proper treatment of connectionism
Paul Smolensky
Department of Computer Science and Institute of Cognitive Science.
University of Colorado, Boulder, Colo. 80309-043https://home.csulb.edu/~cwallis/382/readings/482/smolensky.proper.treat.pdf

|  | Virus-free. www.avast.com  |