The rise of Real AI Industry: Causal Interactive Learning vs. Deep Statistical Learning
AI technology is turning our world upside down, with Emergent AI to Have a Profound Impact on Business and Society [see the 2023 Hype Cycle for Emerging Technologies]
Major technology companies together with start-up businesses face the global challenge, accelerating their digital transformation and harnessing the power of AI to help overcome the present emergency and meet new challenges in a new world.
Today's AI as machine learning techniques, deep learning algorithms and deep neural networks can’t make sense of, understand or interact with the world, namely, to identify its causality, causation or interaction, its elements and structures, processes and mechanisms, rules and relationships, data and models, all what makes our world.
Relying on statistical regularities instead of causality leads to all sorts of decision and prediction errors, data and algorithmic biases, the lack of quality data, and implementation failings, or the absence of real machine intelligence and learning.
In all probability, Causal or True or Real AI has to be created by 2025 years replacing the whole fake AI industry of LLMs and specialized AI chips for training and inference of generative AI algorithms. Now as AI chips are going graphics processing units (GPUs), field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), or Central Processing Units, implemented as a core on system-on-a-chip (SoC). With state-of-the-art fake AI chips, training a fake AI algorithm can cost tens of millions of U.S. dollars and take weeks to complete, and a large portion of total spending is on AI-related computing.
It is not surprising that the U.S., Taiwanese, and South Korean firms controlling chip fabrication factories (“fabs”) fabricating fake AI chips are speculatively blooming.
The Taiwan Semiconductor Manufacturing Company (TSMC)?makes all of the world's narrow/weak AI chips, supplying fabless chipmakers, as Nvidia's A100 GPU; Qualcomm, AMD, Broadcom, it also includes the FAI chips from Google, AMD, Amazon, Microsoft, Cerebras, SambaNova.
Most AI chips in the world today are owned by a few cloud providers, reselling them “as a service.” This includes the cloud giants Amazon Web Services, Microsoft Azure and Google Cloud Platform, as well as Oracle Oracle, CoreWeave?and?Lambda Labs.
Besides, there are many companies readying or shipping chips for Edge AI , including SiMa.ai, Hailo Technologies, AlphaICs, Recogni, EdgeCortix, Flex Logix, Roviero, BrainChip, Syntiant, Untether AI, Expedera, Deep AI, Andes, Plumerai, in addition to Intel, AMD (Xilinx) and NVIDIA.
Many naively believe that the geopolitics of fake AI chips could define the future of AI .
What your read and hear of now is going to its illogical end as a fake AI with a false machine learning, be it large language models, as GPT-3-4-5, or virtual mobile assistants. It is mere a publicity hype, marketing lingo or mass commercial fraud, choose yourself.
All weak/narrow overspecialized AI chips are to be disrupted by general-purpose AI chips, as Causal Neural Networks will replace Numerical Neural networks—the basic algorithmic architecture powering every important FAI breakthrough over the past decade, from AlphaGo to AlphaFold to Midjourney to ChatGPT.
As a result of memory needs, specialized semiconductor AI chips are of large chip size, it is very costly to manufacture a specialized AI chip for every application.
A general-purpose AI platform would help address this challenge . "System and chip vendors would still be able to augment the general-purpose platform with accelerators, sensors, and inputs/outputs. This would allow manufacturers to customize the platform for the different workload requirements of any application while also saving on costs. An additional benefit of a general-purpose AI platform is that it can facilitate faster evolution of an application ecosystem".
A general-purpose true AI will emerge as Interactive Causal AI , what partly has been predicted in the Gartner Hype Cycle for Emerging Technology 2022-2023, as Causal AI .
Below what a good mind should know about a computational modeling of reality and causality and causation and its disruptive emerging technology, Causal Machine Intelligence and Learning or Real/Causal AI .
The key message to deliver is AA/AI Rule for Autonomous Machine Intelligence: "There is no True AI without Understanding the Cause and Effect of Interactions within the World" .
Real AI vs. False AI
Artificial intelligence (AI), machine learning (ML), deep learning (DL), automation and robotics are transforming our world. Understanding the nature of AI/ML/DL is the critical step in building an AI world.
AI must be upgraded as the transdisciplinary science and engineering of making intelligent machines, as complementing and augmenting human intelligence, individual and collective.
The domineering assumption of AI as emulating, mimicking, simulating, or replicating human body/intelligence/brains/mind/behavior is scientifically unjustified and ethically harmful and existentially risky and should be discarded in the favor of non-human machine intelligence and learning as an alternative and augmenting intelligence (AAI).
Towards the next generation AI, Causal Machine Intelligence and Learning
The pandemic turned our world upside down and many countries, major technology companies with start-up businesses face the global challenge, accelerating their digital transformation and harnessing the power of AI to help overcome the present emergency and meet new challenges in a new world.
At its conception, AI researchers attempted to teach, educate, instruct, or program by explicitly coding symbolic knowledge about the world into machines. Today’s AI systems inductively “learn” from selected training data, as from experience, observations and trial and error, as if acquire knowledge on their own, and this is known as machine learning.?
Standard machine and deep learning algorithms extract data patterns as?statistic correlations?from raw data, structured or unstructured. This is true for simple algorithms, like logistic regression, or sophisticated algorithms, like neural networks, which can learn superficial statistical patterns from input data.?
But there is a big BUT, which is a fatal flaw in the whole enterprise. Today's AI as largely machine learning techniques, deep learning algorithms and deep neural networks can’t identify reality and its causality, its elements and structures, processes and mechanisms, rules and relationships, data and models, all what makes our world.
This leads to all sorts of decision and prediction errors, data and algorithmic biases, the lack of quality data, and implementation failings, or the absence of real machine intelligence and learning.
Integrating the symbolic AI with the statistic AI of Machine Learning, Causal Machine Intelligence and Learning makes the next generation of powerful intelligent machines, running the master [causal] learning algorithms. It allows intelligent machines to think about the world, factual and counterfactual, computing its alternatives and scenarios, and effectively interacting with any complex environments, physical or virtual.
Real Artificial Intelligence vs. Fake Artificial Intelligence or Causal AI vs. Acausal ML
As to Forbes' exposition , nine out of every ten machine learning projects in industry?never go beyond an experimental phase and into production. "It turns out there’s a fatal flaw in most companies’ approach to machine learning, the analytical tool of the future:?87% of projects do not get past the experiment phase and so never make it into production .
Why do so many companies, presumably on the basis of rational decisions, limit themselves simply to exploring the potential of machine learning, and even after undertaking large investments, hiring data scientists and investing resources, time and money, fail to take things to the next level?"
There are two big reasons, educational and fundamental. In any ML/DL/AI courses, you’ll be told to learn to program and relearn statistics as though things were starting from scratch, and there were no a multitude of analytics tools, whether it’s the?extremely popular course ?by?Andrew Ng , “Machine learning for average humans ”, “Absolute beginning into machine learning ,” etc.
Now a fundamental reason is causality and causation, causal relationships among causal variables, or complex causal interrelationships instead of confounded spurious correlations. Today's machine learning algorithms are?incapable of identifying causal patterns as they?can only see correlations?which lack any reality or deep sense.
There is increasing number of works trying to show?why Causal AI, a new category of machine intelligence, is the solution. Now, Causal AI is announced as a new category of machine intelligence that understands cause and effect. Leading ML/DL/AI researchers agree that causality is the future of AI.??“Many of us think that?we are missing the basic ingredients ?needed [for true machine intelligence], such as the ability to understand causal relationships in data”, says Yoshua Bengio, a leading figure in DL research.?
Most of them do the same principal mistakes, apply just a linear asymmetrical causality in deterministic or statistical fashions. It commonly suggests that?"the cause is greater than the effect". Then, in reality, small events cause large effects due to the nonlinearity and cyclicity of causation, like as triggering of large amounts of?potential energy, as in a nuclear bomb. Here is a standard obsolete monotonous argumentation of causal fallacy,?false cause, or?non causa pro causa: "correlation does not imply causation" vs. "correlation implies causation".
“Correlation is not causation”. Correlations are symmetrical (if x correlates with y then y correlates with x), they lack direction and they’re quantitative (the “Pearson correlation coefficient”, a standard measure of correlation, is a single number between -1 and 1). In contrast, causes are asymmetrical (if x causes y then y is not a cause of x), directional and qualitative. So correlation and causation are different concepts. Furthermore, causes can’t be reduced to correlations, or to any other statistical relationship. Causality requires a model of the environment. We know this largely thanks to?Turing Award-winning research ?by AI pioneer Judea Pearl”. [CausalLens’ blog].?
In fact, real things are quite opposite. “Correlation implies or suggests causation”. Correlations are symmetrical (if x correlates with y then y correlates with x), as well as causes are symmetrical (if x causes y then y is a cause of x). So, correlation and causation are similar concepts, and therefore causes intuitively reduced to correlations, or to any other statistical relationship. Where there is causation, there is correlation, but also a production or generation from cause to effect, or vice versa, a plausible mechanism, and sometimes common and intermediate causes. Correlation is often used when discoverying causation because it is a necessary condition, it is not a sufficient condition, which is tested by experimentation, as RCTs, or observational studies.
The meta-rule is this: there are common reasons or root causes, (deep, basic, fundamental, underlying, initial or original), call it confounders or whatever else, and one need to look for them in the first place. It is like the monetary or fiscal policy regimes are root causes for historical negative correlations between inflation and unemployment, which have their own root causes.
Overall, there an increasing number of R&D of linear specific/inductive/bottom-up/space-time causality models created in the narrow context of a statistical Narrow AI and ML, as sampled below:
Due to the narrow causal AI assumptions, such causal AI models among other principal things, as generalization and transfer, lack a much deeper research, as it was noted in the article, Towards Causal Representation Learning, namely:
a) Learning Non-Linear Causal Relations at Scale (1) understanding under which conditions nonlinear causal relations can be learned; (2) which training frameworks allow to best exploit the scalability of machine learning approaches; and (3) providing compelling evidence on the advantages over (noncausal) statistical representations in terms of generalization, repurposing, and transfer of causal modules on real-world tasks. b) Learning Causal Variables. causal representation learning, the discovery of high-level causal variables from low-level observations; c) Understanding the Biases of Existing Deep Learning Approaches; d) Learning Causally Correct Models of the World and the Agent. Building a causal description for both a model of the agent and the environment (world models) for robust and versatile model-based reinforcement learning.
All the above are falling in the trap of a linear causality, when a causal effect is defined as below: Causal Effect . If making a change in a quantity X results in a reliable demonstrable change in a quantity Y in a given context, then X has a causal effect on Y.
Nonlinear Causality is Everything for Human Knowledge and Machine Intelligence and Learning
To know the world means to know the causal structure of the world, its mechanisms and causal systems and processes.
Making sense of the world or machine intelligence or human behavior or complex processes as climate change means giving a causal explanation, in the first place, while explaining the world means revealing its underlying causal mechanisms, rules, laws, and invariants.
Real science and technology deal with the cause-and-effect relationships, explaining the internal physical. mental or social mechanisms of phenomena by means of causal laws, rules, and effects.
And mechanisms rule the world, as a network/system of causally interacting parts and processes, individuals, components, activities, or changes, are responsible for producing effects or phenomena of new changes. Mechanisms involve causal interactions of entities and actions/activities, with all their conditions and specifications. It is like natural selection with mutation and gene flow is a mechanism of evolution or predation is a mechanism in ecological systems or physical effects are mechanisms in physical technology, machines, devices and equipment.
Finding and explaining the cause of a phenomenon is to explain the phenomenon and scientists explain phenomena by describing causal mechanisms producing the phenomena.
Formal scientific knowledge corrupted by logical positivism is largely descriptive – telling us?what?and?how, why we?seek explanations, we want to know why, at least, as four fundamental types of answer to the question "why?", the material, the formal, the efficient, or the final.
It is thought that to give a causal explanation is to collect data or give information about causal history.
Overall, any full model of the world or its domain must be descriptive, deductive and inductive, exploratory and explainable, predictive and prescriptive (DDIEEPP),
Besides, the subject of explanation, causality, must be nonlinear causation, with a bidirectional flow of causation between macro and micro levels,?thus
So, the present linear science implies a linear causality, with linear phenomena, while nonlinear science implies a nonlinear causality, with nonlinear structures and phenomena, with all the reality of complex and random processes driven by self-reinforcing or self-regulating causal loops, mechanisms or networks.
The nonlinear causality provides the most plausible models of reality, visualized as causal [loops] diagrams, directed acyclic graphs, linear causal models, or causal Bayesian graphs (expressing the real relationships among random variables in a dataset), all generalized by full causal networks, like as full ANNs.
Nonlinear causality makes all the difference in the next generation of Machine Intelligence, a Causal AI and ML, where causal algorithms could perform description, explanation, exploration, deduction, induction, inference, learning, prediction and prescription in causal networks.
Causal Learning and/or Machine Learning
AI models that could capture causal relationships will be really intelligent and generalizable, unlike ML systems excelling in connecting input data and output predictions, while lacking in reasoning?about cause-effect relations or environment changes.
The most advanced part of ML, Deep Learning (DL), has focused too much on correlation without causation, finding statistic patterns in terms of training data, but failing to explain how they’re connected. The majority of ML/DL successes reduce large scale pattern recognition on the collected independent and identically distributed (i.i.d.) data.?
Causal knowledge and learning are about how intelligent entities think, talk, learn, explain, and understand the world in causal terms, in terms of causes and effects, agents, changes or processes, actions and manipulation.
It is about self-supervised learning, transfer learning and causal discovery, i.e., learning causal information from the real world’s data, from heterogeneous data when the i.i.d. assumption is dropped.
The critical role of causality, causal models, and intervention is evidenced in in the basic cognitive functions: reasoning, judgment, categorization, deductive or inductive inference, language, and learning, and decision making,
Causal learning the cause–effect relationships, as determining the causation among a set of two or more events or discoverying the causality in data,?could be viewed in various ways:
Causal learning of four causes:
Causal learning ?is the process by which people and animals gradually learn to predict the most probable effect for a given cause and to attribute the most probable cause for the events in their environment. Learning causal relationships between the events in our environment and between our own behavior and those events is critical for survival.?
Learning causal relationships can be characterized as a?bottom-up process whereby events that share contingencies become causally related, and/or a top-down process whereby cause–effect relationships may be inferred from observation and empirically tested for its accuracy.
Causal learning ?underpins the development of our concepts and categories, our intuitive theories, and our capacities for planning, imagination, and inference.
All the causal knowledge confusion comes from its defective linear definition, as exposed in the Wiki Article, Causality:
Causality:?influence by which one event, process or state, a cause, contributes to the production of another event, process or state, an effect, where the cause is partly responsible for the effect, and the effect is partly dependent on the cause.
Historically, a faulty linear causality originated from Hume's interpretation of a cause: “we may define a cause to be?an object, followed by another, …where, if the first had not been, the second never had existed” [Hume, 1777, p. 62]. The linear idea of cause here comprises?spatial contiguity?of cause to effect,?temporal precedence?of cause over effect, and?necessary connection?between cause and effect.?
After Hume, the dominant strategy in the analysis of causality has been reductive, linear and statistic or probabilistic, as well as counterfactual, eliminating causality as a basic ontological category.
This all marked by a giant causal fallacy of ignoring the whole world of reverse causality. On an intuitive level, the factorization of the joint distribution P(Cause, Effect) into P(Cause) x P(Effect | Cause) is identical with the factorization into P(Effect) x P(Cause | Effect). It is formulated as Bayes' theorem?(Bayes' law?or?Bayes' rule), making the basis for Bayes inference, statistics and deep machine learning algorithms.
A blue neon sign showing the Bayes's theorem
Such a misleading linear approach leads to systematic biases and errors and reflected in too many studies on causal learning, discovery, inference, modelling, reasoning, etc., including 3 levels of causality, as Association, Intervention, and Counterfactuals (J. Pearl).?
The Real AI’s Ladder of Reality [Causality and Mentality, Science and Technology, Human Intelligence and Non-Human Intelligence (AI)]
In reality, causality and causation are the most complex phenomena in the world, as it is evidenced by the AA ladder of CausalWorld, as in 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 (MI or 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).?
Overall, the Causal Machine Intelligence and Learning involves the deep learning cycle of World [Environments, Domains, Situations], Data [Perception, Percepts, Sensors], Information [Cognition, Memory], Knowledge [Learning, Thinking, Reasoning], Wisdom [Learning, Understanding, Decision], Interaction [Action, Behavior, Actuation, Adaptation, Change], and new World…
Cyclical Nonlinear Causal Orders or Circular Orderings vs. Linear Monotonous Causal Orders
Orders are everywhere in the world, be it hierarchies or categorization. They are studied by mathematics, in order theory, as special types of n-ary relations.
[Causal] Orders are special binary relations of events/changes, which are both symmetric and antisymmetric, or coreflexive, relating every element to itself.
Suppose that P is a set of events and that ≤ is a causal relation on P. Then ≤ is a partial causal order if it is reflexive, antisymmetric, and transitive, that is, if for all a, b and c in P, we have that:
a ≤ a (reflexivity)
if a ≤ b and b ≤ a then a = b (antisymmetry)
if?a?=?b?is true then?b?=?a?is also true (symmetry)
if a ≤ b and b ≤ c then a ≤ c (transitivity).
There are always the converse, dual relation, or transpose, of a causal binary relation, as?the?opposite?or?dual?of the original relation,?or the?inverse?of the original relation,?or the?reciprocal?of the original causal relation.
There are special types of causal order, as partial orders, linear orders, total orders, or chains.?While many familiar orders are linear, real orders are nonlinear, as a?cyclic causal order, a way to arrange a set of things/objects/events in a?circle.
A cyclic causal order on a set?of changes X?with?n?variables or elements is like an arrangement of?X?on a clock face, for an?n-hour clock. Each element?x?in?X?has a "next element" and a "previous element", and taking either successors or predecessors cycles exactly once through the elements as?x(1),?x(2), ...,?x(n).
The general definition is as follows: a cyclic causal order on a set?X?is a relation?C???X, written?[a,?b,?c], that satisfies the following axioms:
领英推荐
And cycles overrule linear orders . Given a cyclically ordered set?(K, [ ])?and a linearly ordered set?(L, <), the (total) lexicographic product is a cyclic order on the?product set?K?×?L, defined by?[(a,?x), (b,?y), (c,?z)]?if one of the following holds:
The lexicographic product?K?×?L?globally looks like?K?and locally looks like?L; it can be thought of as?K?copies of?L. This construction is sometimes used to characterize cyclically ordered groups.
All in all, for any two correlated events, changes or variables, x and y, all ther possible relationships include:
The cyclical causation goes as feedback, which “occurs when outputs of a system are routed back as inputs as part of a chain of cause-and-effect that forms a circuit or loop”. In a feedback loop all outputs of a process are available as causal inputs to the process. There are a positive or negative feedback, named as?self-reinforcing/self-correcting,?reinforcing/balancing,?discrepancy-enhancing/discrepancy-reducing?or?regenerative/degenerative, respectively. Besides, any positive feedback could be a virtuous or vicious cycle, depending on its interaction effects, constructive or destructive.
Feedback is used extensively in physical, chemical, biological, engineering, or digital systems. Feedback can give rise to incredibly complex behaviors, as biological systems contain many types of regulatory circuits, both positive and negative.
Causal feedback systems, mechanisms, loops, and interactions are involved in any complex behavior of any piece of reality, physical, biological, mental, economic, political, ecologic, social, virtual or digital.
Here is an illustrative case study of the most challenging macroeconomic problems, a business cycle recession and recovery.
How Real AI Sees a Human Economy with its Recession and Recovery
A recession is a specific sort of causal vicious cycle, with cascading declines in output, employment, income, and sales that feed back into a further fall in output, spreading rapidly from industry to industry, country to country and region to region.?
On the opposite side, a business cycle recovery begins when that recessionary vicious cycle reverses its causation and effects and becomes a virtuous cycle, with rising output triggering job gains, rising incomes, and increasing sales that feed back into a further rise in output. The recovery can persist and result in a sustained economic expansion only if it becomes self-feeding, which is ensured by this domino effect driving the diffusion of the revival across the economy.?
How the RAI is to Measure Recovery/Recession Business Cycles
The severity of a recession is measured by the three D's: depth, diffusion, and duration. A recession's depth is determined by the magnitude of the peak-to-trough decline in the broad measures of output, employment, income, and sales. Its diffusion is measured by the extent of its spread across economic activities, industries, and geographical regions. Its duration is determined by the time interval between the peak and the trough.?
In analogous fashion, the strength of an expansion is determined by how pronounced, pervasive, and persistent it turns out to be. These three P's correspond to the three D's of recession. Now, let us see how this all could impact in the real life context.
2 Must-Know AI Concepts In Real ML/DL
It is interrelated concepts of causal learning and self-supervised learning, replacing supervised and reinforcement learning.
Supervised deep learning systems ?have dominated the AI landscape since 2012. These systems learn from labeled data to classify new instances into the learned classes.
Reinforcement learning?resembles the way humans learn, while doing a set of actions to achieve a reward. DeepMind researchers even published a paper arguing that “reward is enough ” to achieve general artificial intelligence.
Recently, researchers have recently put more interest in the paradigm of?unsupervised?— or?self-supervised learning. Humans learn a lot by?observing ?and perceiving the world. That’s what self-supervised learning is about.
“[Self-supervised learning] is?the idea of learning to represent the world before learning a task.?This is what babies and animals do. […]?Once we have good representations of the world, learning a task requires few trials and few samples.”
In all, the progress of machine learning systems looks like as the 4-5 steps ladder:
Some successes of the transformer architecture are due to self-supervised learning, like as BERT or GPT-3 used in language generation tasks. Self-supervised systems are now state-of-the-art in many NLP domains. A big drawback of these systems is not their inability to handle continuous input such as images or audio, but missing the key element of machine intelligence - Causal Learning.
Why is Acausal Non-Real Fake AI so harmful?
Five top-performing tech stocks in the market, namely, Facebook, Amazon, Apple, Microsoft, and Alphabet’s Google, FAAMG, represent the U.S.'s Narrow AI technology leaders whose products span standard machine learning and deep learning or data analytics cloud platforms, with mobile and desktop systems, hosting services, online operations, and software products. The five FAAMG companies had a joint market capitalization of around $4.5 trillion a year ago, and now exceed $7.6 trillion, being all within the top 10 companies in the US.
As to the modest Gartner's predictions, the total FAI-derived business value is forecast to reach $3.9 trillion in 2022.
You don’t need to be a great economist to foresee that such a speculative, circular and leveraged mega bubbles lead the global COVID-19 plagued economy to its deep recession and real economy collapse.
A real solution here is not a fake and false, narrow and weak, acausal AI of ML and DL, relying on blind statistics and mathematics to imitate some specific parts of human cognition or intelligent behavior.
What the pandemic-stricken world needs, it is the Real AI Technology which must be developed as a digital general purpose technology, like a Synergetic Cyber-Human Intelligence.
Human minds as collective intelligence and world knowledge will be integrated with the Human-Machine Intelligence and Learning (HMIL) Global Platform, or Global AI:
GAI = HMIL = AI + ML + DL + NLU + 6G+ Bio-, Nano-, Cognitive engineering + Robotics +
SC, QC + the Internet of Everything + Human Minds + MME, BCE + Digital Superintelligence =
Encyclopedic Intelligence = Real AI = Global AI = Global Cyber-Human Supermind
Again, the 4th Industrial Revolution (4IR) as a fusion of advances in artificial intelligence (AI), robotics, the Internet of Things (IoT), genetic engineering, quantum computing, and other digital technologies transforms human economy into machine economy.
A human-like AI/ML/DL technology might rapidly make entire industries obsolete, in either case triggering a widespread mass unemployment, while over-enriching the FAI Big Tech.
Ultra-Large AI Models Are Over: The end of "reward is enough" and “scale is all you need” is near
From 2020 to 2022 — until very recently — most high-repercussion, newsworthy announcements in AI were on LLMs (AlphaFold is one of the few notable exceptions).
It was during this period that phrases like “AGI is coming” and "reward is enough" and “scale is all you need” became super popular.
In the paper "Reward is Enough" its DeepMind's authors "hypothesise that the objective of maximising reward is enough to drive behaviour that exhibits most if not all attributes of intelligence that are studied in natural and artificial intelligence, including knowledge, learning, perception, social intelligence, language and generalisation. This is in contrast to the view that specialised problem formulations are needed for each attribute of intelligence, based on other signals or objectives. The reward-is-enough hypothesis suggests that agents with powerful reinforcement learning algorithms when placed in rich environments with simple rewards could develop the kind of broad, multi-attribute intelligence that constitutes an artificial general intelligence".
Another way to AGI is " scale is all you need". OpenAI set a precedent. Google, Meta, Nvidia, DeepMind, Baidu, Alibaba… the major players in the field lost no time. Their priority was surpassing GPT-3. It wasn’t a competition with OpenAI but an attempt at corroborating the rumors: Did scale work so well? Could AGI really be around the corner?
Big tech companies bought the scale argument and wanted to signal their presence in the AI race. Here’s a brief, incomplete list of how the landscape changed in one year, from mid-2021 to mid-2022 [company: model (size, release date)]:
A pretty dramatic picture — cherry-picked to back my argument, yes, but quite revealing regardless. Companies were running away from small-scale AI.
But, what were they looking for in between hundreds of billions of parameters? They didn’t know.
Scale proved to improve performance, but were benchmark results translatable to real-world performance? They didn’t know.
Could they really reach AGI with sheer size? Could scale alone lead us to intelligence?
They also didn’t know.
Every few months a company released a new largest model. But they were escaping forward from having to think about the limitations. They didn’t have a plan. They didn’t know where they were going or, most importantly,?why.
DeepMind found that all super-large models are “significantly undertrained.” They’re?unnecessarily?big.
Why make models larger when there’s room for improvement at lower sizes?
The human brain has ~100 billion neurons x ~10,000 connections. That’s 1,000 trillion synapses. If we accept the complexity delta of two orders of magnitude, we’d need a model with 100 quadrillion parameters to reach the scale of the human brain.
That’s 500,000 times the largest AI model in existence today.
Maybe pursuing AGI mindlessly through pure scaling isn’t so reasonable after all.
AGI, human-level AI, human-like AI, superintelligence, true AI, strong AI… Imprecise terminology is a symptom we don’t know much about what we’re talking about.
The lack of definitions and measuring tools creates an insurmountable gap between our knowledge of reality and reality itself. [Ultra-Large AI Models Are Over ]
How to Build Real/Causal/General AI
A General Artificial Intelligence (gAI) is computing machinery capable of deep understanding and effective interacting with the world autonomously learning how to carry out a huge range of tasks in a huge range of environments.
GAI is emerging now as Real/Causal AGI, while being featured in science-fiction stories as a human-like and human-replacing AGI for more than a century, and popularized in modern times by sci-fi films.
Fictional depictions of AGI vary widely, tending towards the dystopian vision of intelligent machines eradicating or enslaving humanity, as in The Matrix or The Terminator or I- Robot, or Ex Machina.
In such stories, AGI is often cast as either indifferent to human suffering or bent on mankind's destruction.
This narrative is all about a human-like AGI, which is mimicking, replicating, simulating human body and brain, mind and intelligence and behavior with the goal to replace Homo Sapiens as new Machina Sapiens.
It is one thing to inspire and stimulate, the other one is to copy and simulate.
Many things might inspire your imagination and curiosity and creativity, be it bird's flight or fish's swimming or the human body and the human brain.
But it is a critical mistake to copy or simulate in your inventions all of them, be it bird's flight or fish's swimming or the human body and the human brain.
Again, the bird's flight is about the aerodynamics of flight, the balance of four causal/aerodynamic forces, Lift and Weight, Thrust, and Drag.
There is no magic but the magic of science.
The same refers to the human body, brain, and intelligence. You will certainly fail identifying machine intelligence with human intelligence, as the 70+ human-like AI/ML histories demonstrate it.
The single source of truth, knowledge and intelligence is the world itself , with its causal principles, rules, laws, regularities and models. This is the only rational, scientific way to build machine/computing/technological intelligence, which is to complement the human mind, not to replicate and finally replace humans.
As simple as that.
Conclusion
Today’s Narrow AI of ML/DL/NNs is about spurious statistical learning instead of real causal learning, with all the consequences.
Such a fake or false or pretended AI, presented by machine learning techniques, deep learning algorithms and deep neural networks, can’t identify reality with causality, its elements and structures, processes and mechanisms, rules and relationships, data and models, and how it is reflected in mentality and virtuality. It is all what makes our world.
As a result, the most of ML/DL/NAL “models never hit the market”.
SUPPLEMENT
The 10 Commandments about Interaction, Causality and Causation
Sources
Experienced Data Analyst with detail-oriented analytical skills, experienced in complex dataset analysis, delivering actionable insights to drive business decisions.
1 年this is interesting