Global Trans-AI Platform = RE = Machine World Model [Learning, Inference, Interaction] Engine + AI + ML + LLMs (GPT-4, BERT, T5, Wu Dao, MT-NLG...)

A Reality Engine (RE) is a real AI technology platform that knows or learns, understands and draws inferences, predicts and interacts based on comprehensive scientific world modeling, not merely statistics and correlations, of data input.

Such Machine World Model [Learning, Inference, Interaction] Engine is superior to a traditional machine learning AI platform, generative AI and causality engine.

Why a human-like, human-level AI must be overruled

Starting with Frankenstein, all sci-fi art industry (literature, movies, TV series, videogames, virtual realties, and what not) is promoting a human-like and human-level AI, be it robotic humanoids or alien intelligence.

As evidenced by all sorts of space-wars films, from Dune to Start Wars and Start Trek and Space Odyssey, to urban dystopian sci-fi, Metropolis, the Terminator, the Matrix, Blade Runner, I, Robot, or Ex Machina, such cosmic lies thrill and attract, terrify and fascinate the billions of people, bringing a lot of attention and money to its stakeholders, however dumb, dull and defective the whole idea might be.

By its design and nature, such a sci-fi human-like superintelligence is plagued with all sorts of negative AI disruptions:

• Total job automation, all predictable and repetitive works and knowledge workers positions, from black-collar to golden collar, are susceptible to NWAI takeover
• Privacy, security, and ‘deep fakes’, as “Orwellian” usage of facial recognition technology in open spaces, workplaces and educational institutions, the EU AI Act view today's human-like generative AI as a high-level risk AI , and unacceptable risk AI systems considered a threat to people and will be banned
• AI bias, ML model are unable to solve complex real-world problems due to inherent machine bias and variance
• Widening socioeconomic inequality
• Existential threats

According to Elon Musk, CEO of Tesla and SpaceX, AI is more lethal than nuclear weapons: "Mark my words; AI is far more dangerous than nukes,” he stated. Similar to this, Stephen Hawking warned us all that the impact of AI might be catastrophic if its rapid progress is not tightly and morally regulated. This is all about a human-like and human-level AI, as ANI, AGI or ASI.

Such a concern was documented in an open letter "Pause Giant AI Experiments : We call on all AI labs to immediately pause for at least 6 months the training of AI systems more powerful than GPT-4".

As for the General/Real/True AI/Robotics/Automation, the leader is that country which is the first to embrace the sustainable and smart development idea of Real AI/Robotics/Automation/Economy/Industry/Society.

Global Trans-AI as Transdisciplinary/Translational/Transformative Intelligence

Learning causal world models for effectively interacting with the world is the key feature of any real-world intelligence, animal, human or machine.

The machine world model learning, inference and interaction (WMLII) is powerful and effective ways to create artificial intelligent machines and to understand intelligence in the context of how an intelligence (an AI, robot, or human) interacts with the world at large and their worlds (environments).

The way to transform today's AI, with its ML applications, DL algorithms and data analytics methods, is its integration with comprehensive scientific world models:

Real/Generalized/Global/Transdisciplinary/Translational/Transformative Intelligence = World/Reality/Causality Model Engine + ANI + ML + DL + LLMs + AGI + ASI

Machine-learning AI technologies, such as deep learning and reinforcement learning, can be used for pattern recognition in the real world. However, real intelligence is not just about finding patterns, or statistical regularities in data. Real intelligence is about all the ways to build models of the world [through multimodal perceptual systems]?for learning and inference and interaction.?

What we need is to build computational models for the WMLII in order to build a general-purpose machine intelligence or to understand the human brain/mind/intelligence.?

General causal learning and inference is the essential component of real artificial intelligence running on highly-complex computing, high performance, or high efficiency, computing architectures such as?CPUs, GPUs, NPUs, TPUs, etc. It is achieved through the WMLII machine/engine that applies causal/logical rules to evaluate and analyze new information or?to identify and categorize new things.?

True/Real-World/Global Intelligence (GI) embraces as its parts weak or narrow AI applications, neural networks, machine learning, deep learning, multiple linear regression, RFM modeling, cognitive computing, predictive intelligence/analytics, language models, or knowledge graphs.

Real and True AI [Machine Intelligence and Machine Learning] driven by its Machine World Model Learning, Inference, Interaction Engine is processing, analyzing, classifying and interpreting ML training web data sets and big data sets of sorts and types to extract information and knowledge, data insights or intelligence, all as causal patterns and predictions, decisions and solutions.

Be it any type or sort of big data sets: Customer Data, E-commerce;?Transactions; Financial Transactions; Government and Public Data; Health and Medical Records; Internet of Things Devices; Research and Scientific Data from research experiments, academic studies, scientific observations,?digital twin?simulations, and genomic sequencing; Sensor Networks Data gathered from environmental sensors, industrial machinery, traffic monitoring systems; Social Media Platforms Data generated from?social media platforms?like Quora, Facebook, Twitter, Instagram, and LinkedIn, posts, comments, likes, shares, and user profiles.

Global Trans-AI Paradigm

Global Trans-AI is all about reality and its systems and interactions in terms of all in terms of data types or variables, categorical or numeric, such as mathematical objects and concepts, ontological, scientific and technological concepts, as physical, mental, social, economic, political, or environmental constructs,

Global Trans-AI models learn and infer how this digital universe of encoded data variables is interrelated and interacting, and how its members are referring to senses and references, classes and schemas, decisions and predictions and actions.

Global Trans-AI is about interactive machine intelligence and learning, knowing and understanding, its models, schemas, algorithms, architectures, data, information, knowledge, and how it is all used to understand and interact with the world, its entities and processes.

Global Trans-AI intelligence is in its represented world’s categories, concepts, abstractions, types and tokens, and how they all interrelated following fundamental causal algorithms, laws and rules.

Any formal science, as computer science, mathematics or statistics, is but al tool for truer and more real intelligence and its intelligent applications.

The Global Trans-Intelligence (GI) Formula

W x D x C = RI = GI = Man-Machine Global Intelligence

W - World/Reality [Causal] Knowledge (natural knowledge N (physical, chemical, biological), mental knowledge M, social knowledge S, NL knowledge L, engineering knowledge E)

multiplied by

D - World's Data/Information (Unstructured and Structured Data, the Internet/Web data, Digital Data and Analog Data)

multiplied by

C - World's Computing power (total computing power) or Intelligent Power (total human intelligence)

equals

Real/Global Intelligence (Natural Intelligence, Human Intelligence, Artificial Intelligence, Social Intelligence, Technological/Machine Intelligence, Cyber-Physical Intelligence)

Special cases: Animal Intelligence, Human Intelligence, Artificial Intelligence, Machine Learning, Deep Learning, Machine Intelligence, NLP, Artificial General Intelligence, Artificial Superintelligence

W x D x C > B x C x D = AHH — which means Biological knowledge multiplied by Computing power multiplied by Data equals the Ability to Hack Humans.

The Pentagon has been investigating how to fundamentally alter what it means to be human, funding research into creating super humans that are smarter, faster, and stronger through human performance enhancement.

Large Language Models and GI

The development of large language models has been a major milestone in narrow AI, machine learning and deep learning in recent years.

The new LLMs/NLP models launched by Meta and OpenAI as,?Galactica and ChatGPT , has once again activated AI hype and AI haters.

There are a number of LLMs, as pictured above, and the most prominent include GPT-3, ChatGPT (OpenAI), BERT, T5 (Google) or Wu Dao (Beijing Academy of Artificial Intelligence), MT-NLG (Microsoft), META/Galactica.

The LLMs can give us the impression of being capable of doing a lot of things. Their training from huge web corpora in an unsupervised way with the latest neural network architectures, Transformer Models, consisting of a Deep Learning Model and equipped with an Attention Mechanism. They are able to process sequential input (e.g. text, video, audio) and produce corresponding output, all on the basis of statistics and predictive analytics.

Having a huge ecological footprint, the LLM models were trained with several megabytes or terabytes of publicly available web data such as Wikipedia and similar large corpora. The training content is equivalent in size to that of well over a hundred thousand or million books, while a human could read more than 5,000 books in a lifetime.

This all gives them power in complex NLP applications, text generation, translation, complex question-answering systems, generation of summaries, images, video, audio, etc.

But it is obvious that having been trained only with natural language datasets, it is not possible for the LLMs to acquire complex internal world models regarding modeling reality and causality, inferences and logic, knowledge of the physical models, converting unstructured data into structured causal data and process it all autonomously, automatically and intelligently.

As such, ChatGPT is a hugely expensive language toy application, confusing facts and fictions, requesting a lot of data and resources, computing and financial, with no reasoning or understanding, intelligence or real learning.

It is a statistical learning language machine to globally generate the trash data, fake news, misinformation and disinformation. The same refers to all other LLMs, as GPT-4, BERT, T5, Wu Dao, MT-NLG, Galactica…

From Narrow/Weak/Fake AI to Real Intelligence

There are three kinds of AI emerging as fake/false/imitating AI or Real/Scientific/True AI:

• Narrow/weak human-like superintelligence AI systems (ANI), imitating parts of human intelligence.
• General human-like and human-level AI systems (AGI, Full AI, Strong AI), imitating all human intelligence.
• Really intelligent, autonomous machines, augmenting and complementing humans.

Machines operate differently, in terms of world models and causal patterns, but applying quantities and data, numbers and statistics, figures and digits, tokens and syntax, mathematics and probabilities, precision and accuracy, computing power and algorithms.

It is a stimulus-response black box model, having its multi-inputs and multi-outputs (or transfer characteristics, a transfer function, system function, network function, generalized as a transfer matrix) producing useful conclusions/information without showing any information about its internal workings, which mechanisms/explanations remain opaque/“black.”

Humans think in terms of world models and causal rules, but applying qualities, senses and meanings, concepts and ideas, semantics and pragmatics, biases and prejudices, knowledge and wisdom.

The black box AI refers to the inability to provide understandable explanations for their decisions and actions. This lack of transparency and interpretability can make it difficult to understand how AI systems make decisions, to identify and correct errors or biases in AI systems.

To make sense of AI, it must be explainable, when its machine learning model provides explanations for its predictions and decisions, using techniques such as linear regression, decision trees, or random forests.

Building Real AI as Causal Machine Intelligence and Learning

True and Real AI embraces as its parts data-processing systems: weak or narrow AI applications, neural networks, machine learning, deep learning, multiple linear regression, RFM modeling, cognitive computing, predictive intelligence/analytics, language models, or knowledge graphs.

Be it web searches or self-driving transportation, GPT-3-4-5 or BERT, Microsoft' KG, Google's KG or?Diffbot ,?training their knowledge graph on the entire internet, encoding entities like people, places and objects into nodes, connected to other entities via edges.

Without the Real AI, its world/reality/data modeling and simulation (R&M&S), today's?"AI is meaningless" and "often just a fancy name for a computer program" , software patches, like bugfixes, to legacy software or big databases to improve their functionality,?security, usability, or?performance.

"Such machines are not yet self-aware and they cannot understand context, especially in language. Operationally, too, they are limited by the historical data from which they learn, and restricted to functioning within set parameters" .

If?Diffbot ?aims to "build the world's first complete map of human knowledge to enable intelligent systems, crawling the entire public web and providing the world's largest searchable knowledge graph, it first needs "a theory of reality" (dubbed as the general schema of things) encoded and programmed, like "the theory of mind", genetically encoded or socially cultivated, in human minds.

The Real-World AI means a new reality shift from a dumb and dull statistic data universe to a meaningful world of physical and digital realities, mixed reality, augmented reality, virtual reality, or simulated reality, all running by the Trans-AI entities, machines, robots, engines, codes, algorithms, and platforms.

The Trans-AI General-Purpose Technology is to change the world in all respects and aspects, the economy and industry, society and life, governments and international relations.

Today, even with a statistic AI/ML/DL, every day we are witnessing new kinds of developments in each part of life, from politics to economics. It is impacting our daily life with language machines, deep fake news, chatbots, IoT devices, travel navigation, fraud detection, mass surveillance, smart home devices, smartphones, voice assistants, drones, self-driving cars, cybersecurity, and LAWs.

As Stephen Hawking had noted: “Success in creating effective AI, could be the biggest event in the history of our civilisation. Or the worst. We just don’t know. So we cannot know if we will be infinitely helped by AI, or ignored by it and side-lined, or conceivably destroyed by it”.

In 5 years, we could reach the critical stage of Real AI, a Synthesized Machine Intelligence and Learning (SMIL).

It is to converge most emerging technologies, as

narrow AIs, statistic ML,

rules-base, symbolic AI,

AGI, LLMs,

explainable AI,

Robotics, Smart Automation,

the Internet of Things,

5–6 G,

biometrics,

man-machine interfaces, neurochips,

AR (augmented reality)/VR (virtual reality),

blockchain,

NLP (natural language processing),

and quantum computing.

Conclusion

It is two different worlds, the world of quantitative/physical/cybernetic machines vs. the world of qualitative/emotional/feeling/live humans.

Machines are machines, humans are humans, they can only complement each other, as Real AI, Trans-AI, Causal Machine Intelligence and Learning, or Man-Machine hyperintelligence.

World model learning and inference , Neural Networks , Volume 144 ,?December 2021, Pages 573-590

"World model learning and inference are crucial concepts in brain and cognitive science, as well as in AI and robotics. World model learning is a fundamental mechanism of human and artificial cognitive systems and contributes to a wide range of cognitive capabilities, e.g.,?pattern recognition, action selection,?social cognition , language learning, and reasoning".