Real AI = Machine Intelligence and Learning = Causal AI + Narrow AI + Statistical ML + ANNs
Real AI is not a human-mimicking artificial intelligence (AI) and machine learning (ML), but CAUSAL Machine Intelligence and Learning.
It is emerging as a man-machine hyperintelligence, integrating symbolic rules-based AI (SAI), narrow and weak AI (ANI), statistical ML, and deep neural networks (DNNs) into causal data analytics (CDA), causal AI (CAI), and collective human intelligence (CHI)
RAI = MIL = SAI +CDA + ANI + ML + DNNs + CAI + CHI?
Fake AI: Imitative AI and Statistical Machine Learning
Today's AI is statistical software and hardware with statistically trained competences, skills. or data literacy, imitating the human-like ability to read, understand, reason, create, and communicate?data?as information and knowledge, predictions and decisions. It is aimed to understanding what data means, drawing conclusions from data, producing content of any forms, modalities and styles, as it is pursued by the so-called large language models (LLMs), from GPT to BERT to Wu Dao, "the biggest language A.I. system yet", having 1.75 trillion parameters.
The statistical machines with probabilistic models and algorithms lack deep data understanding due to a lack of understanding of causal relationships, which is the biggest obstacle to giving them real general intelligence and deep causal learning.
This mostly comes from the conceptual fault of all statistical learning systems consisting in its defective world's model, dealing with random samples and probabilities, instead of the world of entities and their interrelationships, meanings and senses.
A random sample is modeled as independent and identically distributed (i.i.d. iid or IID) random variables. It is a set of objects chosen randomly, or, more formally, “a sequence of IID random data points or tokens”. Predicting in the IID settings is defective; for "statistical models are a superficial description of reality as they are only required to model associations".?The majority of current successes of ML due to large scale pattern recognition on suitably collected i.i.d. data, while engineering away interventions in the world, generalizations, domain shifts, and temporal structure.
At best, it is upgraded as the so-called Markov chains?or?Markov processes, discrete or continuous,?describing "a?sequence?of possible events in which the probability of each event depends only on the state attained in the previous event". Here predictions are conditional/contingent?on the present state of the system, or its future and past states are?stochastically/statistically/randomly independent.
The IID or Markov property assumptions are aimed to simplify the underlying mathematics, being completely unrealistic, while driving all the mentioned LLMs, as AI chatbots, from Google/LaMDA/Bard to Microsoft/OpenAI/ChatGPT, changing the writing completely.
Here is a simple and clear description of the imitative, human-like AI, with its branches as machine learning and deep learning.
"Artificial intelligence refers to the ability of machines to perform tasks that would normally require human intelligence to complete. This includes tasks such as recognizing patterns, making decisions, solving problems, and learning from experience. AI systems can be designed to mimic the cognitive processes of humans and can be programmed to perform a wide range of tasks.
The goal of AI is to create machines that can perform tasks as well as or better than humans, and that can continually improve their performance through experience and learning.
Machine learning is a subset of AI that focuses on the development of algorithms and models that enable machines to learn and make predictions based on data. Machine learning algorithms use statistical techniques to enable computers to find patterns in data and to make predictions based on those patterns. There are several different types of machine learning algorithms, including supervised learning, unsupervised learning, and reinforcement learning. These algorithms can be trained on large amounts of data and can continue to improve their accuracy as they are exposed to more data. Supervised learning algorithms are trained on labeled data, and are used to make predictions about future events. Unsupervised learning algorithms are used to discover patterns in data, and are often used for tasks such as clustering and dimensionality reduction. Reinforcement learning algorithms are used for decision-making tasks, and allow systems to learn from experience by exploring different options and receiving feedback on their performance". [How Machine Learning is Different from Artificial Intelligence?]
"State-of-the-art AI is relatively narrow, i.e., trained to perform specific tasks, as opposed to the broad, versatile intelligence allowing humans to adapt to a wide range of environments and develop a rich set of skills. The human ability to discover robust, invariant high-level concepts and abstractions, and to identify causal relationships from observations appears to be one of the key factors allowing for a successful generalization from prior experiences to new, often quite different, “out-of-distribution” settings.
A central problem for AI is...causal representation learning, the discovery of high-level causal variables from low-level observations".?
Causal AI and ML: SOTA AI vs. Causal Inference AI
Causal AI (CAI) is defined as a statistical?AI (SAI)?system that can explain [linear] statistical causality, causation, or cause and effect, and the CAI technology is to help explain decision making and the causes for a decision.
The assumption is “Machines' lack of understanding of causal relations is perhaps the biggest roadblock to giving them human-level intelligence.” [The Book of Why: The New Science of Cause and Effect, 2018]
In reality, it is humans' lack of understanding of causal relationships as interactions or productive correlations or associations is the biggest roadblock to giving machines human-level intelligence.
Correlation?means there is a statistical association between variables.?Causation?means a real association between variables when a change in one variable causes a change in another variable, and VICE VERSA.
Correlation and causation are two entangled things at all levels, from the quantum level to the cosmological level of the whole universe: “correlation implies causation”, while "causation postulates correlation".
First of all, there is one real relationships in the world and reality, which is causation or cause and effect, all the the rest, as resemblance, contiguity, wholeness
Any really good researchers in any domains are implicitly guided by the ten fundamental features of real causation:
production or generation or impact,
correlation or covariation or association,
reversibility or retroactivity or reaction and response,
co-occurrence, contiguity,
preceding causation, the web of causation,
sufficiency and necessity,
interaction, interrelationship, interconnection,
alteration, change,
time order, simultaneity or temporal precedence
causal looping, cycles, feedback loops, positive or negative.
"The cause precedes the effect (time order). The cause co-occurs with the unaffected entity in space and time (co-occurrence). Causes and their effects are the result of a web of causation (preceding causation). The intensity, frequency, and duration of the cause are adequate and the susceptible entity can exhibit the type and magnitude of the effect (sufficiency). The cause effectively interacts with the entity in a way that induces the effect (interaction). And, the entity is changed by the interactions with the cause (alteration)".?
Still, many dogmatically presume that "correlation doesn't imply causation" and that causality is the asymmetrical relations of cause and effect, the cause (independent variable) must precede the effect (dependent variable) in time, as the treatment/intervention 'causes' the outcome in the RCTs (A/B) experiments.
Causation is defectively defined as?the relationships between independent, control, and dependent variables, as a linear, asymmetrical relationship between two variables X and Y such that changes in X lead to changes in Y, and the key difference between association and causation lies in the confounding. The causal metamodel is confused as a three-level abstraction, the ladder of causation, as associations/probabilities/correlations/regularities/patterns (seeing/observing); causality/intervention (doing); counterfactuals/but-for causation/sine qua non?causation/physical laws (thought experiment/imagination/reasoning). [ Pearl, J.?Causality: Models, Reasoning, and Inference. Vol. 64 (Cambridge University Press, 2000)].
This all corrupts the whole science and engineering of causality and causation, the real nature of causal processes and mechanisms, causal analysis (regularity, probabilistic, counterfactual, and manipulative), reasoning and inferences and models, impact, efficacy and effectiveness, measures and effects.
The linearity of causality brings to the poor conception of causal graph or causal network (also known as path diagrams, causal Bayesian networks) as a topologically ordered directed acyclic graph, DAG, with no directed cycles, where one event (X) is to cause another if it raises the probability of the other (Y).
In statistics, econometrics, epidemiology, genetics and related disciplines, causal graphs as DAGs are?probabilistic graphical models encoding assumptions about the data-generating process. Such statistics revolves around the analysis of relationships/associations/correlations among multiple variables.
In reality, causal structure or causal system is the network graph of causal interrelationships [associations/correlations/interactions] among or between a set of causal variables, and causal discovery algorithms are the problem of learning the causal structure, patterns and correlations, rules and laws, from data, observational and experimental, simulation or theoretical, real-world or synthetic, structured or unstructured, digital or analogue.?Real machine statistics revolves around the analysis of causal relationships among multiple variables/nominal, ordinal, interval, ratio, numeral data.
Gartner has included causal AI in its 2022?Hype Cycle?report, citing it as one of five critical technologies in accelerated AI automation.
The idea of causal/physical/real AI was first generated by us, in 1989, as part the Encyclopedic AI project (A. Sh. Abdullaev, Preprint Knowledge Base of Encyclopedic Artificial Intelligence, Moscow, 1989, Academy of Sciences of the USSR).
Among the developers of causal AI software are causaLens, Xplain Data,?Geminos?and Qualcomm.
Causal AI autonomously searches for causes in data, while also boosting experimentation and human intuition. Causal AI discovers the underlying causal relationships in data, whereas machine learning just analyses correlations.?
Causal knowledge can be acquired in three mutually complementary ways:
through?experimentation; via?human expertise?and intuitions; with?causal discovery?algorithms
"Causal AI frameworks and algorithms support decision making tasks like estimating the impact of interventions, counterfactual reasoning and repurposing previously gained knowledge on other domains".?
State-of-the-Art AI vs. Causal AI
Decision-making AI
Causal AI doesn’t just predict the future, it shapes it.
Explainable AI
Put the “cause” in “because” with next-generation explainable AI.
Adaptable AI
Causal AI continuously adapts to real-world dynamics.?
Human-centric AI
Human-plus-Causal AI partnership allows organizations to harness the benefits of AI.?
Imaginative AI
Causal AI can explore hypothetical worlds, uncovering insights that explain why events happened.
Fair AI
AI has a bias problem and Causal AI is the solution.
领英推荐
AI for small data
70% of organizations are shifting their focus from big to small data — Causal AI can help.?
Trustworthy AI
Trust is the most important but often-overlooked ingredient in successful AI adoption.?
How are Causal AI models different from Bayesian networks?
The two types of models have some superficial similarities, but they also have significant differences. Bayesian networks (BNs) simply describe patterns of correlations between variables. Causal AI models capture the underlying processes that drive those statistical relationships.
This paradigm shift makes Causal AI models more flexible, versatile, and powerful than Bayesian networks.
In fact, a?Bayesian network?(a?Bayes network,?Bayes net,?belief network, or?decision network) is a probabilistic?graphical model?that represents a set of variables and their?conditional dependencies?via a?directed acyclic graph?(DAG), where efficient algorithms can perform?linear causal inference and learning.
What we really need is realistic generalizations of Bayesian networks, in which not only?probabilistic inference or decision making?problems (following the?optimum of objective, loss or reward functions, as the maximum expected utility?criterion) can be modeled and solved.?
Real AI: Man-Machine Intelligence and Learning (MMIL)
Real AI Engineering: Reality > Causality = Interaction > Real Science and Technology > Artificial Intelligence and Machine Learning > Statistical AI/ML/DL > Causal AI > Interactive AI = Real Machine Intelligence and Learning > Trans-AI
Unlike a narrow and weak AI and causal AI, Real AI is a generalized intelligence, integrating symbolic AI, statistical AI/ML/DL and causal AI with human intelligence.
RAI is driven by its Global Intelligence Engine, the World Hypergraph: Global Causal Graph Network: Interactive Bayesian Graph Networks, governed by the?general chain rule?(the?general product rule), where all the world's variables (causal factors) are interacting (conditionally dependent of one another) with each other.
It underlies all the possible architectures of artificial neural networks, as the most advanced deep neural networks algorithms.
RAI has knowledge, competence, skill or literacy in ANY field of knowledge or practice:?science literacy; technology literacy; computer literacy;?statistical or data literacy;?critical literacy;?media literacy;?ecological literacy,?disaster literacy;?health literacy;?social literacy;?quantitative literacy (numeracy),?visual literacy, e.g. body language, pictures, maps, and video), etc.
The RAI Enginery includes:
The foundations of RAI:
Philosophy/Metaphysics/Ontology/Epistemology/Logics/Semantics [Reality/Reason/Knowledge/Meaning]
Science/Physical/Biological/Social
Mathematics and Statistics and Data Science
Computer Science and Information Science and Engineering
Cognitive Science, Psychology and Linguistics
Engineering and Technology, Automation and Robotics
Business Intelligence…
Real AI is the telos of humanity
The only real future of homo sapiens is homo-machina sapiens.
There are no real AI entities in existence, yet. But what is certain, such a technological hyperintelligence will be like creating an alien ETI intelligence.
As such, it will make
the discovery of all discoveries,
the innovation of all innovations,
the invention of all inventions,
the machine of all machines,
the technology of all technologies
Real AI or really intelligent autonomous machine is emerging as the greatest scientific discovery and engineering innovation of all time, with which no scientific or technological developments could compare, be it
Fire
Agriculture
Language
Transportation
Urbanization
Electricity
Gravity, Theory of relativity, Quantum theory
Evolution, DNA
Periodic Table
Computer
Internet
Space exploration…
Again, Real AI should not be confused with an imitative AI, as making software/hardware thinking intelligently, in a similar way the intelligent humans think, perceive, understand, predict, learn, decide and work and manipulate a world.
It is NOT about implementing human intelligence in machines i.e., creating software/hardware systems that understand, think, learn, and behave like humans, sold as the “Grand AI Dream”, “Strong AI” or “Human-level and Human-like AI”.
Again, the original goal of AI, is not a human-imitative Arti?cial General Intelligence (AGI) aiming at the development of “thinking machines”, general- purpose systems with an intelligence comparable to human intelligence.
Certainly, it is not what the mainstream of imitative AI research has been focused on: domain-dependent and problem-speci?c solutions, as the weak and narrow AI of machine learning models and deep learning methods, be it applied as self-driving cars or LLMs, as LaMDA/Bard or ChatGPT.
Without question, the most groundbreaking achievements in all human history is Real Human-AI Superintelligence and invention of really intelligent hybrid cyber-physical socio-technological systems, software and hardware, platforms and networks, completely upending the common order of things, disrupting all the human conditions.
SUPPLEMENT on Causal Data Analytics
CDA converts raw data into the causal data of actionable insights. It includes a number of types, tools, technologies, and processes used to forecast, find trends as objective patterns and solve real-world problems by using real-world data. CDA can shape business processes, decision-making, and business growth.
It covers the 5 types of data analytics: