I would rather discover one true cause than gain the kingdom of Persia. Democritus
EIS Encyclopedic Intelligent Systems has developed the first real model of the Real/Causal AI, including the following elements of its Universal Intelligent Platform, I-World:
Machine World Model;
Master Algorithm, Causal.World;
World Data Framework, World.Data;
Global Knowledge Base, World.Net;
Domain Knowledge Base, Domain.Net
The development has reached a stage of a proof of principle, concept and mechanism in which the best AI technology stack of hardware, software and dataware is constructed and tested to explore and demonstrate the feasibility of the Real/Causal AI Model.
Real AI builds the comprehensive causal models of environments, represented/codified/operationalized as the world's data universe. It creates a world-data mapping of all possible entities, their relationships and behaviors, binding causes and effects.
To build truly AI machines of infinitely powerful digital intelligence, we need to encode, program or teach them what the world is with all its complex cause-effect relationships. What makes machine intelligence and learning a true and real AI is the powerful underlying causal master algorithms used to reveal the causal patterns in the world's data universe.
Free-model and black-box AI systems as deep reinforcement learning agents are inherently unable to learn intelligent behaviors without having to learn a a true causal model of how the world works.
Designed to make sense of the whole of reality, Causal AI or Real Machine Intelligence is emerging as the Next Generation of Artificial Intelligence and Machine Learning. But it is still a very novel concept for the key stakeholders, researchers, developers and engineers; policy makers and funders; program managers and experts of any sectors where causal AI should be applied.
AI is the hardest ever challenge for the human mind.
It combines all the best sciences and technologies as philosophy, ontology, logic, mathematics, computer science, cybernetics, physics, biology, cognitive science, robotics, engineering together.
AI has all potential to revolutionize everything, economy, industry, government, society, cities, all sides of human life.
AI is to change completely human history and the order of things we used to have for thousands years.
AI is emerging as a radically new class of life and intelligence with cosmic implications.
Creating a real causal AI/MI/ML/DL which is powerful to decide the most complex world's problems, as Leibniz's Superscientist (LS) and Laplace's Demon (LD), could be the most critical general purpose technology.
The Next Big Thing is emerging as an integrative general purpose technology (IGPT) for all emerging technologies helping humanity in solving major world problems, from the crisis of leadership to public health global issues to climate change. Here is the hierarchy of the next big things in technology:
1. Artificial Intelligence (AI) as Machine Intelligence and Learning (IGPT)
2. Autonomous driving
3. Reusable rockets
4. Virtual Reality and Augmented Reality
5. Renewable energy technologies at a global scale
6. Superfast AI internet of things
7. Edge Computing
8. Quantum Computing
9. 6G AI
10. AI Cyber Security
11. Online DNA analysis
12. Immune system engineering...
The first candidate technology is the machine intelligence and learning GPT, as a global Human-AI Platform. We are at the brink of solving the biggest problem in human history, a techno-mind at a global scale.
To build such hybrid man-AI machines of infinite digital intelligence, we have to teach them what the world is with its cause and effect. What makes machine intelligence and learning a true and real AI is the powerful underlying causal master algorithms used to reveal the causal patterns and regularities in the world's data universe.
You might think of "Infinite Man-Machine Intelligence" as Skynet from the Terminator or the Architect from the Matrix movies.
Ultimately, there might be two kinds of intelligence:
We traditionally define the three broad categories of human-like AI:
So, the difference between an Human Intelligence, ANI, AGI and an ASI is that the latter, virtually, will never exist in a physical form, as being housed in any super powerful quantum computing hardware.
ASI is never limited by the energy resources, the hardware or software or algorithms and programs or data.
Such infinite intelligence operates entirely within a supercomputer or global network of supercomputers, having full access to all data stored on the Internet, as well as whatever device, emerging technology or human that feeds data into and over the Internet.
This means there will be no practical limit to how much, how fast, how far or how deep the ASI can learn or self-improve or pervade or impact.
Causality is the source and essence of all reality, the soul of the whole world. Its causation is interrelating the things in the world, allowing us to make predictions about the future, explain the past, act on the present, and intervene to change outcomes, to intelligently shape human environments.
The nature of causality is still a big unknown unknown regardless of it being systematically investigated in many academic disciplines, including philosophy and physics, biology and social sciences, statistics, probability theory, and computer science [Causal Relationship].
As a result, in academia, there are a significant number of theories on causality, physical, biological, mental, social, or informational.
In physics, the cause and the effect are connected through a local mechanism (impact) or a nonlocal mechanism (field), in accordance with known laws of nature.
In causal data science, "data causality" or "causal discovery" makes an exploratory causal analysis, an extension of exploratory data analysis, "the use of statistical algorithms to infer associations in observed data sets that are potentially causal under strict assumptions".
Some social science researchers see causation as not just a correlation, but a counterfactual dependence to manipulate the world, conceptually or experimentally, as a randomized controlled/clinical/comparative trial or intervention study vs. observational study. Randomized experiments are used for the estimation of population-level causal effects while comparing two futures, the "treatment" and the "control".
Causal modelling as describing the causal mechanisms of all sorts of real-world systems could have a deciding role in AI and ML and DNNs.
The state-of-the art AI is largely big data-driven, but meaningful data analysis is primarily concerned with causal analysis and causal models and estimating causal effects, all pertaining to establishing cause and effect.
Presently, there is a big hype around AI as Machine Learning and Deep Learning, focusing on finding correlations in data sets, classifying data and predicting. It is all instead of understanding and explaining the world, its true causality, with all descriptions, discovering, explanations, prediction, prescriptions or interventions.
ML/DL is great at finding correlations in data, but not real causation, falling into the trap of equating correlation with causation.
Other crucial widespread error is failing to distinguish Ontological/Generic/Deductive/Top-Down Causality vs. Phenomenological Specific/Inductive/ Bottom-Up/Space-Time Causality.
Causality is about a generic causal relation, with a generic set of causes and effects, with the structural invariance to individual objects or systems, space and time. It is as different as ontology vs phenomenology, theoretical science vs. empirical science, as observables, an appearance, action, change and occurrence of any kind vs. the force/the law producing it.
There are an increasing number of research and development of naive, linear specific/inductive/bottom-up/space-time causality in terms of a statistical AI and ML, as sampled below:
All of them are confined by a critical misunderstanding of causation that "if X causes Y that does not mean that Y causes X". Then the reverse phenomena is the whole disruption of causality and causation, thus enabling all kinds of interactions of all kinds of real world systems and processes: "if X causes Y that DOES mean that Y causes X".
Again, one of the key formal constructs, Bayes' principle, was invented by its author as being about a reversible generic probabilistic causation. And Bayesian causal networks are used only in studies of general causation, operationalized as singular causality.
"Smoking causes lung cancer" makes most valuable causal knowledge, and it is not "Somebody's smoking causes his lung cancer".
Such an unexampled cognitive bias of naive linear, one-sided causation essentially limits our ability to rely on our causal models and theories as well as on our advanced intelligent systems, AI, ML and DL, for deciding complex world's problems, from climate change to human-like AI robots.
Causality is considered to be a major principle, a universal law or a conditio sine qua non of human practice, science and technology.
The first principle which human knowledge, as philosophy and science, might formulate is the principle of causality, which still reigns supreme. The essence of standard causality is the one-sided generation and determination of one phenomenon by another.
The standard Causality Principle states that everything necessarily has a cause. The principle indicates the existence of a relationship between two things/events, the cause and the effect, and an order between them: the cause always precedes the effect. The principle of causality has been variously stated in the history of philosophy and science: Every effect has a cause. Every contingent being has a cause. Whatever comes to be has a cause. What is, has sufficient reason for its existing. The Principle of Sufficient Reason states that everything must have a reason, cause, or ground.
Dictionary Definition of linear causality and causation
Causality is the relation of cause and effect. Causality is causal action or agency. Causality is the doctrine or principle of causes.
Causation is the act, process or agency of causing, producing or effecting; the act or agency which produces an effect. Causation is the relation of cause and effect.
“Correlation does not prove causality.”
[Linear] Causality (also referred to as causation, or cause and effect) is influence by which one event, process, state or object (a cause) contributes to the production of another event, process, state or object (an effect) where the cause is partly responsible for the effect, and the effect is partly dependent on the cause.
[Linear] causality concerns relationships where a change in one variable necessarily results in a change in another variable. [Linear] causation is the demonstration of how one variable influences (or the effect of a variable) another variable or other variables. There are three conditions for causality: covariation, temporal precedence, and control for “third variables.”
[Real] Causality or true causation, or a general cause-effect relationship is an interaction (mutual or reciprocal or circular process, action or influence) by which one general change (a class of events, actions, processes as a cause) produces another general change (a class of events, actions, or processes as an effect) where the cause is productive of the effect, and the effect is inversely productive of the cause.
What is generally missing is the basic truth: Real Causality is the Mother Law, the Law of Laws, the Principle of Principles, or only universal law, being symmetrical, reversible, invertible, convertible and interactive, reversing in effect, action, influence sequence, direction, position, order, or condition, to interact with its cause.
Real Causality or True Causation is the Principle of Principles, Laws of Laws, as well as the Master AI Algorithm.
The Real Causality Principle states that all phenomena in the world are causally interrelated or interacting.
Nowhere in the world can there be any phenomena that have not been caused by other phenomena, and vice versa.
The Principle of Real Causality (PRC) states that causality is a genetic interrelationship of phenomena through which one thing (the cause) under certain conditions causes something else (the effect), and vice versa.
Or, all causal relations are inverted, and all causal laws are symmetric. The mechanism of real causality is associated with the integrated transference of matter, energy and information, as cyberspace. a) physical infrastructures and telecommunications devices; b) computer systems (see point a) and the related (sometimes embedded) software that guarantee the domain's basic operational functioning and connectivity; c) networks between computer systems; d) networks of networks that connect computer systems (the distinction between networks and networks of networks is mainly organizational); e) the access nodes of users and intermediaries routing nodes; f) constituent data (or resident data).
Again, real causality is about a generic causal relation, with a generic set of causes and effects, with the structural invariance to individual objects or systems, space and time.
And any association or correlation or interrelationship implies causality; correlation is an essential component of causality, as a base to determine all causality, linear and nonlinear. Real life and science and technology is all about cause-and-effect relationships, their natural laws, mechanisms, systems and processes, following the Master Algorithm of Science and Intelligence, human and artificial:
Real Causation/Universal Causality (Causal World/Causal Knowledge/Science & Technology) =
Interrelationship (Correlation, Associations, Laws, Rules, Algorithms, Structural Models, Causal Model of the World, Bayesian causal networks, World's Data Patterns, Theoretical Science, Computational Modeling and Simulation) + Observation (Empirical science, Statistics) + Manipulation or Intervention [Experimental science, Experimentation: Hypothesis testing, H0 (null hypothesis) or H1 (primary hypothesis), or A/B/n experiments]
A causal relationship to exist between X and Y IFF there is a bivariate association between X and Y.
The Master Algorithm of Causal Discovery is Descriptive, Deductive, Intuitive, Inductive, Exploratory, Explainable, Predictive and Prescriptive (DDIIEEPP).
It underpins the scientific method of inquiry depending on the reciprocal interactions of phenomena observations [questions] and generalizations, making conjectures (hypotheses) by inductions and testing hypotheses by conducting experiments or studies, deriving predictions as logical consequences from them by deductions, and then carrying out experiments or empirical observations based on those predictions. Such principles of the scientific method, testing whether the real world behaves as predicted by the hypothesis by an experiment, observational study, field study, or simulation, underlie the development of science since at least the 17th century.
Again, the Master Algorithm covers five contemporary machine-learning paradigms—evolutionary algorithms, connectionism and neural networks, symbolism, Bayes networks, and analogical reasoning—unified in one “master algorithm” potentially capable of learning nearly anything.
Domingos' envision of machine learning methodologies into a “master algorithm.” The Master Algorithm, Pedro Domingos.
In physics, the principle of causal interactionism is operationalized as the fundamental interactions (fundamental forces) that form the basis of all known interactions in nature: gravitational, electromagnetic, strong nuclear, and weak nuclear forces, allowing a fifth force as well to explain various anomalous observations, like dark matter and dark energy, that do not fit existing theories.
Production or generation, determination, symmetry or covariance, reflexivity and transitivity are the postulata of Real Causality and True Causation.
The Mass-Energy-Information Trinity conservation might come from the universal causal symmetry defining a symmetry (or covariance) of the fundamental laws of nature/physics and the equivalence between energy and mass.
If Causal AI is to evaluate whether or not some X causes some Y, it needs to launch 5-step causal algorithms:
(1) what classes, kinds or types of observables or causal data or variables or changes are involved, independent, dependent, confounding, controlled, etc.?
(2) is there covariation/correlation/link between X and Y?
(3) is there a control for all confounding variables Z that might make the relationship between X and Y a spurious false association or statistic dependency [the Common Cause Principle, the researcher are commonly “failed to control for” some potentially important cause of the dependent variable]?
(4) what is first observed, if X causes Y? Or if Y causes X? Or which is a reversed causation? Or what is a causal effect?
(4) determine causation, test for causation, direct and reversed, find a correlation, anf test for causation by running experimentation “controlling the other variables and measure the difference or causal effect”: Causation = Correlation + Observation + Manipulation or Intervention [experimentation: Hypothesis testing, H0 (null hypothesis) or H1 (primary hypothesis), or A/B/n experiments]
(5) what is a comprehensive nonlinear circular causal model/mechanism/process/path/channel that interrelates X with Y [the “how” and “why” questions]? In the context of AI, it is causal perception-action loops, the cornerstone of autonomous systems such as self-driving vehicles.
The same process can be applied to a wide variety of causal claims and questions of different complexity. Does drinking red or white wine cause a reduction or increase in heart disease? Does psychotherapy help people with emotional and relational problems? Do increases in government spending boost or retard economic growth? Does democratization cause economic development? Do parental effects (Z) make the relationship between pre-school experience and later educational or learning outcomes spurious? What are the “causes” of historical and social events (the French Revolution, or World War I, the Great Depression, World War I, the Cold War, the COVID-19)?
The Causal AI is analyze the public data sets taking the following programming steps:
Finding appropriate empirical evidence to evaluate the degree to which a causal claim or theory is or is not supported.
Developing a causal theory about the relationship between an independent variable (X) and a dependent variable (Y) from the data set.
Identifying the credible causal mechanism that connects X to Y.
Explaining its answers: (b) Could Y cause X? (c) What other variables (Z) would you like to control for in your tests of this theory?
In each of these and many other examples, we need to observe a correlation between variables before concluding that the relationship is causal, like as "if economic development causes democratization, then democratization causes economic development".
Spotting, identifying and measuring causal claims make an essential general learning skill, with a causal decision-tree scorecard as a shorthand for summarizing the answers to these 5 questions in square brackets [Yes/No, Yes/No, Yes/No, Yes/No, Yes/No].
In all, the real causation paradigm organizes all possible relations, from dependence, association, correlation, or statistic relationship to a naive, linear causation to a circular nonlinear causality, as the AA ladder of CausalWorld:
The Six Layer Causal Hierarchy defines the Ladder of Reality, Causality and Mentality, Science and Technology, Human Intelligence and Non-Human Intelligence (AI).
The CausalWorld [levels of causation] is a basis for all real world constructs, as power, force and interactions, agents and substances, states and conditions and situations, events, actions and changes, processes and relations; causality and causation, causal models, causal systems, causal processes, causal mechanisms, causal patterns, causal data or information, causal codes, programs, algorithms, causal analysis, causal reasoning, causal inference, or causal graphs (path diagrams, causal Bayesian networks or DAGs).
It reviews a causal graph analysis having a critical importance in data science and data-generated processes, medical and social research and public policy evaluation, statistics, econometrics, epidemiology, genetics and related disciplines.
The CausalWorld model covers Pearl's statistic linear causal metamodel, as the ladder of causation: Association (seeing/observing), entails the sensing of regularities or patterns in the input data, expressed as correlations; Intervention (doing), predicts the effects of deliberate actions, expressed as causal relationships; Counterfactuals, involves constructing a theory of (part of) the world that explains why specific actions have specific effects and what happens in the absence of such actions.
It must be noted that any causal inference statistics or AI models relying on "the ladder of causality" [The Book of Why: The New Science of Cause and Effect] are fundamentally defective for missing the key levels of real nonlinear causality of the Six Layer Causal Hierarchy.
Gottfried Leibniz (1646-1716), a giant in mathematics, logic and philosophy, imagined a scientist who could see the events of all times, just as all times are thought to be present to the mind of God.
"Everything proceeds mathematically...if someone could have a sufficient insight into the inner parts of things, and in addition had remembrance and intelligence enough to consider all the circumstances and take them into account, he would be a prophet and see the future in the present as in a mirror."
In the introduction to his 1814 Essai philosophique sur les probabilités, Pierre Simon Laplace, a giant in mathematics, physics, and astronomy, extended an idea of Leibniz which became famous as Laplace's Demon
"We may regard the present state of the universe as the effect of its past and the cause of its future. An intellect which at a certain moment would know all forces that set nature in motion, and all positions of all items of which nature is composed, if this intellect were also vast enough to submit these data to analysis, it would embrace in a single formula the movements of the greatest bodies of the universe and those of the tiniest atom; for such an intellect nothing would be uncertain and the future just like the past would be present before its eyes."
A Laplace Demon has a critical application in the digital age of infinite data. The Data Era is mostly about the world’s digital data, the result of our increasingly digitised way of life.
It is now all about zettabytes information streaming flows, and how to intelligently process it for insights, predictions, recommendations and actions.
As such, there are about M = 8 billion of data units, as individuals, and hundreds of data features, F, as biometric, psychological, behavioral, social, economic, political or geographic data items.
A global human population data ocean, H=M x F, could have billion data records of each unit's attributes,age, gender, race, religion, political party, education, geography, job, income, interest, sentiments, health, etc.
Now, add up here engineered items, billions of roads, buildings, machines, computers, smart devices, weapons, etc., with a view to become digital members of the global zettabyte internet of things.
Now, as the amount of data in the world has increased exponentially, how we are going to govern all of it without global datasphere machine intelligence, its analytic tools, codes, algorithms, programs, applications and robots.
Again, according to market intelligence company IDC’s ‘Data Age 2025’ paper, if the ‘Global Datasphere’ (the total of all data created, captured or replicated) in 2018 reached 18 zettabytes, in 2025 the world’s data will grow to 175 zettabytes.
Add up Cerebras Wafer Scale Engine 2, the largest data processing chip ever built. The Cerebras WSE-2 powers the revolutionary CS-2 system. 2.6 Trillion transistors and 850,000 AI-optimized, fully programmable cores – all packed onto a single silicon wafer to deliver world-leading AI compute density at unprecedented low latencies.
Data, big data, global data, enormous compute, fast memory and communication bandwidth..., no way forward without the Real AI capable to autonomously transform the world's data into the world's knowledge.
Again, the Laplace's Demon starts coming in with Bayesian Machine Learning as well as with High Frequency Trading (HFT), involving the biggest hedge funds manipulating $100 trillion. HFT is an automated trading platform that large investment banks, hedge funds, and institutional investors employ using powerful computers to transact a large number of orders at extremely high speeds. The algorithmic trading systems use complex algorithms to analyze the markets to spot emerging trends in mls, mks, or nanoseconds. High-frequency traders earn their money on any imbalance between supply and demand, using arbitrage and speed to their advantage.
Today, algorithms address many of the problems and decisions that have long been central to the business of trading. What instrument(s) should be invested in or traded? What price should be bid or offered? What order size is optimal? What should be the response to a request for a quotation? What risk will be taken on by facilitating a trade? How does that risk change with the size of the trade? Is the risk of a trade appropriate to a firm’s available capital? What is the relationship between the price of different but related securities or financial products? To what market should an order be sent? Is it more effective to provide liquidity or demand liquidity? Should an order be displayed or non-displayed? To which broker should an order be sent? When should an order be submitted to a trading center? In general, algorithms utilize a rich array of market information to very quickly assess the state of the market, to determine when, where, and how to trade, and then to implement the resulting trading decision(s) in the marketplace.
Hedge funds over the past ten years have evolved from going after the time dimension (HFT) to working in the data dimension (a Laplace Demon). A Laplace Demon can predict anything as it knows everything instantly and understand their causality links.
AI is powering a dramatic change in every parts of human life, in every society, economy and industry across the globe, emerging as a strategic general purpose technology.
Then it is most significant to know what real and true intelligence is, with all its major kinds, technologies, systems and applications, and how it is all interrelated with reality and causality, and superintelligence, as condensed in the Leibniz's Superscientist and Laplace Demon concept.
What is the Laplace's Demon/Superintelligence?
The Laplace's Demon/Superintelligence is dealing with the universe or reality in terms of the causal world models and data/information/knowledge representations for cognition and reasoning, understanding and learning, problem-solving, predictions and decision-making, and interacting with the environment.
The Laplace's Demon is a man-machine superintelligence or supermind which is modeling and simulating the world to effectively and sustainably interact with any environments, physical, natural, mental, social, digital or virtual. This is a common definition covering any intelligent systems of any complexity, human, machine, or alien intelligences.
Its heart and soul consists in the consistent, coherent and comprehensive causal model of the world.
It is a common knowledge that current ML technology fails when applied to dynamic, complex systems. It produces static models that overfit to yesterday’s world. The models are data-hungry and unintelligible to humans, as listed below:
And Causal AI comes with the competitive features
Causal AI is a new category of intelligent machines that understand cause and effect ― a major step towards true AI.
Teaching machines to know/understand "why" is to transfer their causal data patterns [knowledge, rules, laws, generalizations] to other environments.
True causality knowledge, a deep understanding of cause and effect, is critical for a Causal AI/ML/DL, like in:
A causal AI platform, keeping the advantages of comprehensive digitization and automation – one of the key benefits of machine learning – allowing zetabytes of datasets to be cleaned, sorted and monitored simultaneously, is to combine all this data with causal data models and causal/explainable insights – traditionally the sole competency of domain experts.
Global Wireless RAI = IoT + 5G + AI/ML/DL.
It is a new computing paradigm: a globally distributed ML/AI/DL learning over wireless networks, scaling costly very computationally intensive training beyond the cloud.
AI, the Internet of Things and 5G are improving each over with enhanced service quality, simplified deployment, higher network efficiency, and improved network security; increased radio awareness, enhanced device experience, improved system performance, and better radio security.
As to Qualcomm Technologies Blog: 5G+AI: The ingredients fueling tomorrow’s tech innovations
A central or edge cloud sends a state-of-art global AI model to the devices. Next, each device collects personal data and performs on-device training. On-device AI capabilities increase exponentially, along with improvements in algorithms and software.
On-device ML training has three very important benefits that will lead to mass adoption of AI:
The LS or LD as a causal AI is powerful not only to explain, but to predict, forecast or estimate [causal] timelines of events, human life, societies, nature and the universe.
Today, future is largely unpredictable, we are even unaware of the number of epidemic ways to come.
On a small practical scale, we have predictive analytics with its machine deep learning models, exploiting pattern recognition, to analyze current and historical facts to make predictions about future.
As parts of predictive techniques, there are a lot of forecasting methods, qualitative and quantitative, added up with strategic foresight or futures [futures studies, futures research or futurology] with no big utility. It all revolves around pattern-based understanding of past and present, thus and to explore the possibility of future events and trends.
On a large scale, the nature timeline is a big mystery even afterwards, not mentioning the universe timeline, where we have poor ideas about its history, far future and ultimate fate.
Real [Causal] AI as a Synthesized Man-Machine Intelligence and Learning (MIL) is one of the greatest strategic innovations in all human history. It is fast emerging as an integrating general purpose technology (GPT) embracing all the traditional GPTs, as electricity, computing, and the internet/WWW, as well as the emerging technologies, Big Data, Cloud and Edge Computing, ML, DL, Robotics, Smart Automation, the Internet of Things, biometrics, AR (augmented reality)/VR (virtual reality), blockchain, NLP (natural language processing), quantum computing, 5-6G, bio-, neuro-, nano-, cognitive and social networks technologies.
But today's narrow, weak and automated AI of Machine Learning and Deep Learning, as implementing human brains/mind/intelligence in machines that sense, understand, think, learn, and behave like humans, is an existential threat to the human race by its anthropic conception and technology, strategy and policy.
The Real AI is to merge Artificial Intelligence (Weak AI, General AI, Strong AI and ASI) and Machine Learning (Supervised learning, Unsupervised learning, Reinforcement learning or Lifelong learning) as the most disruptive technologies for creating real-world man-machine super-intelligent systems.