General AI and ML = Trans-AI = Unified World Model Machine + Intelligent Neural Networks
Artificial intelligence is the philosophy and metaphysics, logic and mathematics, science and engineering of making intelligent technologies, machines and computer programs, applying world models, data and causal algorithms, to effectively and sustainably interact with the world.
It is shown that the [conceptual, mathematical and scientific] modelling of reality in terms of the global causal graph networks, or the world hypergraph coded as the world model machine, world knowledge engine and world data engine, forms the basis for General Machine Intelligence and Learning, as a transdisciplinary AI (Trans-AI) overruling a human-like and human-level AI, as narrow and weak AI (ANI), ML/DL, ANNs, AGI and ASI.
It is as the Maxwell–Heaviside equations form the foundation of?classical electromagnetism, classical?optics, and?electric circuits,?providing a mathematical model for electric, optical, and radio technologies, such as power generation, electric motors,?wireless?communication, lenses, radar, electronic devices, computers, etc.
Trans-AI is the fundamental and universal technology for all the general-purpose technologies, digital applications and emerging technologies.
WHY MACHINE INTELLIGENCE SHOULD SIMULATE REALITY INSTEAD OF HUMAN INTELLIGENCE AND/OR BEHAVIOR
From one side, humans have long been designated as the most intelligent species on the planet, with big brains, cognitive abilities and processing power outcompeting all other species.
From other side, there is no fundamental idea or life-critical practice which is without deep defects and gaps, mistakes and confusions, ending with all sort of sins and evils, from poverty to wars.
Regardless of our great intelligent deeds, humans are deeply ignorant of primary things, failing to recognize the depth of ignorance.
Sill we don't know what is reality, being, or the universe is and how the world works,
Still we don't know what relationships are and how correlations and associations, causation and interaction are interrelated,
Still we don't know what is brain/mind/intelligence and how human or artificial intelligence works.
To survive and prosper, humanity needs powerful superintelligent partners, call it artificial intelligence or technological intelligence, computing intellect, cybernetic mind, general machine intelligence and learning, universal intelligence, trans-human intelligence, or Trans-AI [Trans-AI: How to Build True AI or Real Machine Intelligence and Learning].
For due to all the inherent biases and prejudices, hyper-specialization and in-group thinking of human intelligence, building general AI and ML as non-human machine intelligence is a matter of life or death for humanity.
We are totally biased creatures preferring or disliking someone or something more than someone or something else, in the most harmful ways.
We are born, then educated and then misinformed to be biased against everything and everybody else.
Our political systems are biased against true policy
Our social systems are biased against true society
Our economic systems are biased against true economy
Our justice systems are biased against true justice
Our scientific systems are biased against nature
Our technological systems are biased against humans
And all citizens are biased against each other, and so on and on.
THE FIRST TRANS-AI MODEL FOR NARROW AI, ML, DL, AGI, ASI, AND HUMAN INTELLIGENCE
The Trans-AI Science and Technology is set to change how our world works and humans live, study, work, and play.
The Trans-AI, as General Machine Intelligence and Learning is the main engine of the new total digital revolution, scientific and technological, social and economic, cultural and religious.
The COVID-19 crisis has accelerated the need for human-machine digital intelligent platforms facilitating new knowledge, competences and workforce skills, advanced cognitive, scientific, technological, and engineering, social, and emotional skills.
In the AI and Robotics era, there is a high demand for the scientific knowledge, digital competence, and high-technology training in a range of innovative areas of exponential technologies, such as artificial intelligence, machine learning and robotics, data science and big data, cloud and edge computing, the Internet of Thing, 5G, cybersecurity and digital reality.
The combined value – to society and industry – of digital transformation across industries could be greater than $100 trillion over the next 10 years. “Combinatorial” effects of AI, ML, DL, Robotics with mobile, cloud, sensors, and analytics among others – are accelerating progress exponentially, but the full potential will not be achieved without the collaboration between humans and machines.
General Machine Intelligence and Learning without Turing's Prejudice
Today’s AI/ML/DL is not a human-like AI (or Artificial Human Intelligence, AHI), be it virtual assistants or autonomous vehicles or predictive applications or large language models or search engines or recommendation systems or language translators or facial recognition systems or q/a systems or gamers, as Apple’s Siri and Microsoft’s Cortana, IBM Watson or MuZero or ChatGPT-3.
We are fooled and duped by a big tech commercial/advertising/media hype that has a commercial purpose to earn some megaprofit.
Today's AI is mostly driven by statistical learning models and algorithms, designated as data analytics, machine learning, artificial neural networks or deep learning. It is implemented as a combination of IT infrastructure (ML platform, algorithms, data, compute) and development stack (from libraries to languages, IDE, workflow and visualisation).
In all, it involves:
Besides, it is “narrow AI” which is designed to perform a single task, and any knowledge gained from performing that task will not automatically be applied to other tasks.
Today’s machine intelligence and learning could go as an extension of cloud services consumption models:
Or, it is created as a data science applications stack.
Or, it is a Qualcomm-like AI stack, supporting a wide variety of use cases including:
Most AI applications in use today can be categorized as being narrow AI, referred to as?weak AI.
They all are missing general artificial intelligence and machine learning, which is defined by three critical interacting engines:
General AI and ML and DL programs/machines/systems are distinguished by understanding the world as multiple plausible world state representations with its Reality Machine and Global Knowledge Engine and World Data Engine.
It is the most essential component of General/Real AI Stack, interacting with its real-world Data Engine, and providing the intelligent functions/capacities:
So, the narrow/weak AI is still not true, real or general MIL. As a human-like intelligence, it is rather a fake and false AI of MIL, being a computing, digital or techno-intelligence in its own.
AHI is plastered, or AI-washed everywhere. A lot of things branded as a human-mimicking “AI”:
“AI photography,”
“deepfakes AI”,
“AI gaming,”
“AI painting”,
"large language models", from GPT-3 to ChatGPT,
“AI chips”…
What do we have is NOT AHI, Being stochastic, statistic, probabilistic data processing software, it makes a special kind of intelligence, narrow machine intelligence standing in need of the Reality Engine and Data Intelligence Engine.
As such, the dominant work of AI/ML/DL today, if it simply pursues its present course, will never reach true intelligence, which includes things such as an ability for a computer system to causally interact with its reality/settings/environment.
It all comes from the fundamental Turing’s prejudice that AI happens when a man-made machine starts to acquire the ability to “think” and act like a human with human-like intelligence and behavior. It is a fundamentally wrong way to building a true autonomous machine intelligence by mimicking the human brain and brains, cognition and intelligence, learning and behavior. ?Yann LeCun, Chief AI Scientist at Meta, one of the 2018 Turing award winners, and one of the most well-known researchers within the field of ML/DL, as a bright example of this prejudice.
In his recent programmatic paper, “A Path Towards Autonomous Machine Intelligence”, LeCun hypothesizes that the difference of humans (and animals) with machines is that the first ones have the innate ability to learn?world models, or “internal models of how the world works”. His human self-supervised learning brain model included six modules: configurator, perception, world model, cost, actor, short-term memory, operating with world models in the perception-action causal loops, as figured below.
Humans are naturally biased creatures and hardly could make a good model for the real AI, which is without Turing's Prejudice, "machines thinking like humans".
"Prejudice is a preconceived?classification?of another person based on that person's perceived?political affiliation,?sex,?gender,?gender identity,?beliefs,?values,?social class,?age,?disability,?religion,?sexuality,?race,?ethnicity,?language,?nationality,?culture,?complexion,?beauty,?height,?body weight,?occupation,?wealth,?education,?criminality,?sport-team affiliation,?music tastes?or other personal characteristics", described as categorical variables.
Machine Learning?bias, or algorithmic bias or AI bias, refers to the tendency of algorithms to reflect human biases, "systematic patterns of deviation from norm and/or rationality in judgment".
The problem arises when an?algorithm?delivers systematically biased results due to erroneous assumptions of the Machine Learning process. In today’s world, this becomes even more problematic because more and more organizations are adopting AI models and systems.
Algorithmic bias?describes systematic and repeatable?errors?in a computer system that create "unfair" outcomes, such as "privileging" one category over another, as observed in?search engine results?and?social media platforms. The possible sources of bias in algorithms could be: human decision biases, unbalanced training data, differential feature use, proxy variables.
General AI and ML as Transdisciplinary AI
Will a real, non-human general AI and ML be a thing in the close future?
I strongly believe that a fully functional prototype of General AI and ML must be developed by 2025, as it was predicted by E. Musk.
It is quite possible and really probable, but as Transdisciplinary AI (Trans-AI, Meta-AI or Real AI or Interactive AI or Causal MIL), which is modelled and designed to integrate and transcend all AI/ML technologies, techniques, models and methods, which separately never reach even an autonomous machine intelligence and learning (MIL) level:
data analytics,
predictive modeling,
data mining,?
pattern matching,
pattern recognition,
question answering,
self-aware systems,
pattern?recognition,?
knowledge representation,?
automatic?reasoning,?
deep?learning,?
expert?systems,
information extraction,
text mining,
natural language processing,
NLU/NLG
problem solving,
intelligent agents,
logic programming,
machine learning,
deep learning,
artificial neural networks,
machine perception,
artificial vision,
computational discovery,
领英推荐
computational creativity
computational statistics...
Artificial Narrow Intelligence
Artificial General Intelligence
Artificial Superintelligence...
We have as many AI models as its developers and stakeholders: the Big Tech AIs, as Google AI, Meta AI, Amazon AI, OpenAI, Microsoft AI, Tesla AI, Tencent AI, Huawei AI, etc., all sorts of academic AIs, added up with startups AIs and national AI initiatives.
I argue for a transdisciplinary AI or metascientific AI, denominated as AA AI, created as the Superhuman Machine Intelligence and Learning (SMIL) model, integrating the data-driven narrow and weak AI/ML/DL models.
If you decompose any today's AI/ML/DL system, it will look just like bits and bytes, neural networks, lots of data, and mathematical algorithms.
If you decompose the human brain, 86B+neurons, connected to form neural pathways, neural circuits, and elaborate network systems, and structured as the cerebrum, the brainstem and the cerebellum, it is just a bag of neurons, firing electrochemical pathways.
Now to make AI/ML/DL/NNs real intelligent, it must be capable to learn to understand the world, as the abstract world models, space-time representations of the environment, with the power to relate the data/information/knowledge to the objective references, thus generating senses and meanings and values, deep and wide knowledge and intelligence and supremely intelligent actions.
Why Artificial human Intelligence (AI) and human-mimicking Machine Learning (ML) are as the dearest myths, worst misconceptions and largest commercial fraud ever
I have been asking again and again: why so far, nobody has developed any version of AI or ML or DL that has reached the level of a cat, dog, or parrot...
Good news.
AI/ML/DL will never reach the level of any animal intelligence, not mentioning human intelligence.
Bad news. It will outperform all collective human intelligence, as its specialized applications easily outsmart humans in specialized tasks, from chess playing to art painting.
How so?
AI/ML/DL is a hype/buzz word of commercial lingo, if you wish, a commercial fraud on a large scale.
To make things just and fair, for the European Commission, I have suggested to launch a multibillion collective action case vs. the biggest tech corporations exploiting the AI fraud schema, known as G-MAFIA and BAD-Triada, for a misrepresentation or deception which is (1) intentional and (2) designed for a tangible gain.
The?total market caps of the largest tech companies (851) approach to $17.822 T?and most of them, from Apple to eBay, misrepresent themselves as dealing with a human-like AI technology.
I label it as a fake and false AI, because as machine learning and deep learning it has nothing to do with simulating/mimicking/emulating the human brain/intelligence/cognition/behavior.
It is all machine ontology (causal world models), real science and engineering, mathematics and statistics and probability, programming and computing. That's it. No human-like perceptions, concepts, ideas, and thoughts, cognition, reasoning and understanding, intellect, intelligence, sentience and consciousness, decision-making, willing, action and interaction.
It is all entities and things, observations, facts, cases and events, modelled with numbers and tokens, data and statistics, quantities and values, structures and functions and measurement scales (nominal, ordinal, interval, ration, cardinal), digital numbers of zeroes and ones, with all its possible combinations and transformations and operations and causal patterns .
It tokenizes and digitizes and processes all the data universe [to ASCII characters], thus creating [non-human] intelligent computing systems, from self-driving cars to large natural language models.
But it is an intelligence, a new type of intelligence which is to be recognized, like it or not.
It is a digital intelligence, alien intelligence, data intelligence, electronic intelligence, statistical learning intelligence or simply machine intelligence which is orthogonal to human intelligence and its poor imitation, dubbed as AI.
I tried my best to formulate this all as 5 simple AA/AI meta rules somehow to clear the whole messy situation:
AA/AI rule 1: Without understanding the cause and effect of interactions within the world, no AI model, algorithm, technique, application, or technology is real and true
AA/AI rule 2: All Artificial Intelligence is Artificial Human Intelligence (AHI), divided as Narrow AI, Artificial General Intelligence (AGI), or Artificial Superintelligence (ASI)
AA/AI rule 3: Real Machine Intelligence (MI) is NOT Artificial Human Intelligence (AHI), the capability of a computer system to mimic human intelligence/cognitive functions/behavior such as perception, learning, reasoning and problem-solving or NL communication.
AA/AI rule 4: Machine Learning is NOT AHI
Now to the point of the matter, how to engineer the Man-Machine Superintelligence model, designated as Trans-AI or Meta-AI.
Neural Networks Without Human-Like Intelligence
Over the past 60+ years there have been many approaches and attempts to get artificial systems to learn to understand its surroundings and learn from its experiences: decision trees, association rules, artificial neural networks, deep learning, inductive logic, support vector machines, clustering, similarity and metric learning including nearest-neighbor approaches, Bayesian networks, reinforcement learning, genetic algorithms and related evolutionary computing approaches, rules-based machine learning, learning classifier systems, sparse dictionary approaches, etc.
Deep learning architectures (deep neural networks, deep belief networks, recurrent neural networks and convolutional neural networks) are universal functional approximators for unstructured tasks: computer vision, speech recognition, natural language processing, audio recognition, social network filtering, machine translation, bioinformatics, drug design, medical image analysis, material inspection, board game programs, etc.
For each task - a special NNS, unlike human brains, thus making an?Artificial Neural Network Zoo.
To be real intelligent, ANNs need to be upgraded from statistical, non-causal learning NNs to causal learning NNs. Together with emerging quantum neural networks, quantum computer with classical data, classical computer with quantum data, and quantum computer with quantum data, they are specifications of causally interactive neural networks (CINNs), a part of global causal nets (GCNs). The GCNs is a collection of interrelated units or nodes called?causal variables, which embraces biological or artificial neurons. The CINNs constitute the "brains" for the General Machine Intelligence and Deep Learning.
Like as the Human Minds, Deep Learning is Hierarchically Representing Things and Relations (objects, images, audio, text, ...) on the Multiple Layers of Abstraction, using multiple layers to progressively extract higher level features from the raw input, as lower layers of edges to higher layers of concepts, digits or letters or faces.
To make DL NNs real intelligent, they must be capable to learn to effective and efficiently understand the world, as the space-time representations of the environment, the power to relate the data/information/knowledge to their objective references, thus generating senses and meanings and values, knowledge and understanding, insight and wisdom.
See for more detail [Data Ontology for Data Science, Technology and Engineering: AI, ML, Deep Learning, Knowledge Graph, and Smart Web Search]
Encoding a unified world model of how the world works, Trans AI is to organize the world’s data, information, knowledge and intelligence and make it all universally accessible and valuable.?
The Poverty of Human-Level and Human-Like AI, HLHLAI, AGI, Strong AI, Full AI...
The whole idea?to imbue machines with HUMAN intelligence has been doomed from its very inception. It is as possible as Perpetuum Mobile.
Human intelligence is only the subject, not the object of study.
From the 70+ AI history, it is plain and clear that nobody on planet is able to develop AGI, not mentioning ASI, due to its fundamental misconceptions relying on the modeling, mimicking, replicating, simulating or emulating the human brain or mind, cognition, intelligence or human behavior.???
It refers to the big tech companies stuck with statistical ML/Narrow AI, as presented by OpenAI GPT-3 or DeepMind Gato or a series of Large Language Models (LLMs).
The hi-tech companies such as OpenAI and DeepMind are fully focused on robust AI systems with AGI but the tech market considers that they are not capable of AGI. They may be so determined on AGI that they are killing the hopes of experiencing AGI.
Both OpenAI and DeepMind are working on the AGI through GPT-3 and Gato for a long period of time. But these companies are not capable of addressing the first problem in solving complex problems of AGI that includes artificial intelligence models learning new things without any training data.
With the right scientific, objective, reality-centered AI/MIL paradigm, it is possible, plausible and feasible to develop autonomous machine intelligence and learning. But it is all about machine ontology and real-world science and technology, mathematical sciences and statistics, programming and computing.
Its design, development and deployment pipeline is the causal chain from Reality and Interactive Causality to Computational Modeling/Representation to Real-World Applications:
General Machine Intelligence and Learning (GMIL) > World [Intelligence, Learning and Inference] Machine: Reality/Causality > Machine Ontology/Global Data Model > Science/Engineering > Mathematics/Set Theory/Optimization/Calculus/Linear Algebra/Probability > Statistics/Data Science > Programming/Algorithms/ANI/ML/DL/Neural Networks > Causal Regression > Software/Hardware > Real-World Applications > ,,,Man-Machine Superintelligence
Again, human intelligence is just the subject, but never the object of study or experimenting.
Real AI is not mere about computing tools or computational devices as GPUs, field-programmable gate arrays (FPGAs), and application-specific integrated circuits (ASICs), designed for either training or inference ONLY for special tasks, or data analytics techniques, as datasets, statistical learning, neural networks, ML algorithms or language models.
It transgresses its mathematical, statistical, logical and computational tools and techniques with a unified world model and simulation engine.
Encoding a 'unified world model' of how the world works, Trans AI is to organize the world’s data, information, knowledge and intelligence and make it all universally accessible and valuable.?
We are at the edge of colossal changes. This is a critical moment of historical choice and opportunity. It could be the best 5 years ahead of us that we have ever had in human history or one of the worst, because we have all the power, technology and knowledge to create the most fundamental generalpurpose technology (GPT), which could completely upend the whole human history. The most important GPTs were fire, the wheel, language, writing, the printing press, the steam engine, electric power, information and telecommunications technology, all to be topped by real artificial intelligence technology. Our study refers to Why and How the Real Machine Intelligence or True AI or Real Superintelligence (RSI) could be designed and developed, deployed and distributed in the next 5 years. The whole idea of RSI took about three decades in three phases. The first conceptual model of TransAI was published in 1989. It covered all possible physical phenomena, effects and processes. The more extended model of Real AI was developed in 1999. A complete theory of superintelligence, with its reality model, global knowledge base, NL programing language, and master algorithm, was presented in 2008.
The RSI project has been finally completed in 2020, with some key findings and discoveries being published on the EU AI Alliance/Futurium site in 20+ articles. The RSI features a unifying World Metamodel (Global Ontology), with a General Intelligence Framework (Master Algorithm), Standard Data Type Hierarchy, NL Programming Language, to effectively interact with the world by intelligent processing of its data, from the web data to the real-world data. The basic results with technical specifications, classifications, formulas, algorithms, designs and patterns, were kept as a trade secret and documented as the Corporate Confidential Report: How to Engineer Man-Machine Superintelligence 2025.
As a member of EU AI Alliance, the author has proposed the Man-Machine RSI Platform as a key part of Transnational EU-Russia Project. To shape a smart and sustainable future, the world should invest into the RSI Science and Technology, for the Trans-AI paradigm is the way to an inclusive, instrumented, interconnected and intelligent world.
Trans-AI or Meta-AI running a Unified World Model Engine undergrids all the human-like AI models mimicking human brains/intelligence/cognition/mind/behavior, as Meta AI + Google AI + Transformers NNs+ Composite AI.
Conclusion
The mathematical and scientific modelling of the world of reality in the context of interactive causal networks forms the basis for General AI and Machine Learning, defined as a transdisciplinary AI (Trans-AI) overruling a human-like and human-level AI and beyond, as narrow and weak AI (ANI), AGI and ASI.
What is as like as the Maxwell–Heaviside equations form the foundation of?classical electromagnetism, classical?optics, and?electric circuits,?providing a mathematical model for electric, optical, and radio technologies, such as power generation, electric motors,?wireless?communication, lenses, radar etc.
SUPPLEMENT I
There is one absolute truth in the world:
" ALL THE WORLD IS INTERACTION", "all reality is interaction", "everything is interaction", at all its levels, scales and scopes.
At the bottom of the things is not a thing, but the interactions of things. Nothing exists alone, be it an elementary particle or the universe, all exists in interacting with each other, being the product of its interactions. All existences are interaction effects, particles, atoms, molecules, bodies, cells, organs, organisms, groups, societies, planets, stars, galaxies, the cosmos...the universe and beyond.
The "Interaction World" (I-World) master principle/paradigm/model/philosophy/science is the way to the most disruptive discoveries about the world, its elements and components. structure and function, history, origin, dynamics and future.
The very idea of the interactive reality opens its deepest ever secret, how to create a real-world superintelligence (RWS)
Interactive Reality > The World Hypergraph: Global Causal Graph Network: Reality > Interaction > Causality > Ontology > Mathematics > Science > Technology > AI > ML > ANNs > DL > Causal AI >Interactive AI > Meta-AI = Trans-AI = Real AI
SUPPLEMENT II
[N]: AI System
[D]: An AI system is a machine-based system that is capable of influencing the environment by producing an output (predictions, recommendations or decisions) for a given set of objectives.?
[D]:?Artificial intelligence (AI), also known as machine intelligence, is a branch of computer science that focuses on building and managing technology that can learn to autonomously make decisions and carry out actions on behalf of a human being. AI is ... an umbrella term that includes any type of software or hardware component that supports machine learning, computer vision, natural language understanding, natural language generation, natural language processing and robotics. Today’s AI uses conventional?CMOS?hardware and the same basic algorithmic functions that drive traditional software. Future generations of AI are expected to inspire new types of brain-inspired circuits and architectures that can make data-driven decisions faster and more accurately than a human being can.
[Conception]: 'artificial intelligence system' means: ...software that is developed with [specific] techniques and approaches [listed in Annex 1] and can, for a given set of human-defined objectives, generate outputs such as content, predictions, recommendations, or decisions influencing the environments they interact with. The notion of 'AI system' would refer to a range of software-based technologies that encompasses 'machine learning', 'logic and knowledge-based' systems, and 'statistical' approaches: 'machine learning approaches', including supervised, unsupervised and reinforcement learning, using a wide variety of methods including deep learning; 'logic and knowledge-based approaches', including knowledge representation, inductive (logic) programming, knowledge bases, inference and deductive engines, (symbolic) reasoning and expert systems; and, 'statistical approaches', Bayesian estimation, search and optimisation methods.
[Classification]: Machine Learning, NLP, Expert Systems, Computer Vision, Robotics, Planning
AI is including the following technologies:
[Conceptualization]: It uses machine and/or human-based data and inputs to (i) perceive real and/or virtual environments; (ii) abstract these perceptions into models through analysis in an automated manner (e.g., with machine learning), or manually; and (iii) use model inference to formulate options for outcomes. AI systems are designed to operate with varying levels of autonomy.
[Inferences and Predictions]: AI is a general-purpose technology that has the potential to improve the welfare and well-being of people, to contribute to positive
sustainable global economic activity, to increase innovation and productivity, and to help respond to key global challenges. It is deployed in many sectors ranging from production, finance and transport to healthcare and security.
Alongside benefits, AI also raises challenges for our societies and economies, notably regarding economic shifts and inequalities, competition, transitions in the labour market, and implications for democracy and human rights.
Professional Web Designer & Developer | UX/UI Expert | Webflow Specialist | Branding & Design Strategist
2 年Great post! Thank you for sharing, Azamat Abdoullaev