Machine Metaphysics for MI and ML: from universal metaphysics to universal AI technology

To build hyperintelligent machines, teach them the world ontology or how to interact with the world, transforming deep neural networks into causal world hypergraph networks underlying all the real-world networks such as computer networks, natural networks, biological networks, technological networks, brain networks, climate networks or social networks.

Dr. Azamat Abdoullaev

Metaphysics historically emerged as "the first science and prime philosophy", including ontology, universal science and natural theology, since the times of Aristotle.

With times, the first science had been fragmented into special schools and contradictory movements, ending with “The Elimination of Metaphysics Through Logical Analysis of Language”, 1931, R. Carnap, and that fundamental ontology, as a variant of metaphysics, is meaningless, or devoid of sense. As a result, for many people, it is massively prejudiced as "an abstract theory with no basis in reality" thus prejudicing our understanding of the whole world, at large and in details.

In reality, metaphysics is everywhere, from our ordinary life and thinking and concepts to science and technology to culture and religion to economics, politics and society.

With the emerging knowledge technology of advanced artificial intelligence and machine learning, as large language foundation models, operating with petabytes of the internet data, there is a great necessity to reanimate the first science as the universal ontological models of world's entities/relationships/categories/data/information/knowledge.

Universal metaphysics as machine metaphysics is the world model engine of advanced artificial intelligence systems designed to understand, learn and?interact with the world, its models and representations, data/information/ knowledge universes.

As such, universal metaphysics?studies the fundamental nature of?reality and causality, the first causes of things and the?first principles?of:?being or existence,?entity, relationship and interaction, substance and property, identity and change,?cause and effect, space?and?time, actuality and potentiality, necessity and possibility, mind and matter, knowledge and data, nature and society, science and technology.

The task of universal metaphysics is to lay out a complete and coherent, consistent and comprehensive ontology, as the universal model of the world of entities and interactions in any of the special inquiries—whether empirical or mathematical, observational or computational, modal or moral.

Then universal metaphysics/ontology as the causal world ontology is all you need for having a powerful intelligence, human or machine.

Now, machine metaphysics (MM) as the digital world ontology for machine intelligence and learning is all what we need to create real intelligent systems and structures, platforms and networks.

Machine metaphysics/ontology is the application of ontological and mathematical methods to automatically learn and infer from world's data and make predictions and explanations and causal discoveries to effectively interact with the world.

To build true intelligent machines, teach them the world ontology or how to interact with the world, transforming deep neural networks into the causal world hypergraph networks extended as all the real-world networks such as?computer networks,?natural networks, biological networks, technological networks,?brain networks,?climate networks?or?social networks.

Universal Metaphysics as making sense of the world for universal AI technology

Any mind or intelligence. human or machine, understands the world via ontology, either as universal metaphysics or applied ontology, telling us what the world is and how it works and changes, all in the forms of the mental models, worldviews, cognitive schemas, theories, conceptual models, beliefs, or commonsense knowledge.

As the science of being or reality or existence, it has been around with humans for a couple of thousands years, with its fundamental questions.

What is the world or reality at large and how it could be systematically ordered by the highest kinds, ontological categories and classes and individuals?

How to identify its entities and relationships, properties and structures, as the patterns of interactions or the invariants of relationships.

How to see the difference between concrete and abstract things, particulars and universals, possibilities and necessities.

How to interrelate various realities, as ontological reality, physical reality, mental reality, social reality, digital reality, virtual reality and information reality.

How to define special classes, as engineering systems and structures,?machines,?materials, devices,?systems,?processes, and?organizations.?

Or, how to interrelate entities, concepts and data or information, with all possible structures, material and abstract structures, as machines and biological organisms, data structures and conceptual models..

In a deep sense, all our life and our sciences and technologies are about applied ontology and special forms of causality, causation, or interaction, be it natural sciences or mathematical sciences, information sciences or computer science or?engineering..

Now why today's AI as ML and DL, particularly generative AI and LLMs models, are defective by default, being without any real intelligence and learning.

AI models have no encoded or programmed?a?priori?knowledge of the world of reality and existence, causality and mentality, humanity and values,?knowledge and data, programming and computing, etc.?

It is just data pattern?matching,?rote memory, probabilistic combinatorics and statistical associations, with no idea?of world ontology, or how the world is structured, behaves and changes.

Universal Metaphysics and Universal AI Technology

As defined, universal metaphysics?studies the fundamental nature of?reality and causality, the first causes of things and the?first principles?of:?...data and knowledge, science and technology.

Here are McKinsey Technology Trends Outlook 2023, with the 15 tech trends, as pictured above, all grouped into five broader categories, all in silos:

the AI revolution,

building the digital future,

compute and connectivity frontiers,

cutting-edge engineering, and a sustainable world.

Or, there is the Gartner Hype Cycle of Emerging Technologies 2023, all in silos, as pictured below:

Or, here are a large number of independent AI applications, all in silos

Universal metaphysics via general-purpose AI technology is breaking down technological silos integrating all the fragmented digital and emerging technologies.

All in all, universal metaphysics is to break down all sorts and forms of silos, as isolated and fragmented, without any interactions and interrelations, interlinks or interconnections:

mental silos,

scientific silos,

technological silos,

economic silos,

business silos, organization, information silos, or silos mentality,

political silos,

social silos,

individual silos,

cultural silos,

religious silos,

information silos.

Conclusion

If machine learning is the application of statistical methods to automatically learn from data, machine metaphysics is the application of ontological and mathematical methods to automatically learn from data and make predictions and explanations and causal discoveries to effectively interact with the world.

THE INTERSECTION OF PHILOSOPHY AND TECHNOLOGY: EXPLORING MACHINE METAPHYSICS AND THE FUTURE OF AI

Reality, Universal Ontology and Knowledge Systems: Toward the Intelligent World

Supplement

Let me note that "universal metaphysics", "universal ontology", as " machine metaphysics" and "machine ontology" look as new constructs, our original innovations, a sort of conceptual trademarks, if to trust the google search engine,

There is “Metaphysic.AI,?industry leaders in using generative AI and machine learning to create photorealistic Hollywood-quality content, combined with their ethics-first approach and thought leadership they unlock an incredible opportunity for the entertainment industry and beyond”. Some rare articles mentioning "machine metaphysics" in the headlines in different contexts, without any definitions, as

Handcraft and Machine Metaphysics

Machine Metaphysics and the Cult of Techno-Transcendentalism

techno-transcendentalism

It emanates from a group of heavily overlapping techno-utopian movements that can be summarized under the acronym?TESCREAL:?transhumanism,?extropianism,?singularitarianism,?cosmism, the?rationality?community,?effective?altruism, and?longtermism.

Universal Metaphysics V1: Primary Course

Universal Metaphysics

• UM is a pursuit of a white-box understanding of our intelligence and language. If successful, it would help us create more interpretable and safer AI agents.
• We believe that UM, as a roadmap of the mind, if developed, will also serve well as the roadmap towards AGI.
• We hypothesize that there are limits on what is understandable and describable by our mind and language. As a roadmap for the mind, UM aims at probing such limits. This is contrary to the traditional optimistic view that AI systems should be good at solving potentially all problems possible. One interesting task would be to identify what is solvable and what is not.?

Note the principal difference between our Universal Metaphysics and Metaphysics, as defined by Wikipedia:

Metaphysics is the branch of philosophy that examines the fundamental nature of reality, including the relationship between mind and matter, between substance and attribute, and between potentiality and actuality. Metaphysics studies questions related to what it is for something to exist and what types of existence there are. Metaphysics seeks to answer, in an abstract and fully general manner, the questions: What is there? What is it like? Topics of metaphysical investigation include existence, objects and their properties, space and time, cause and effect, and possibility. Wikipedia