“Knowing is not enough; we must apply. Willing is not enough; we must do”

– Johann Wolfgang von Goethe

 PROLOG

In our prior article we discussed complexity – the knowns and the unknowns, and how data can be stored in multiple different schema representations to make sense of the information so that it can be rendered meaningfully. In this article on our series on Everything Connected, we merge the dual relationship between “thinking” and “doing” into our construct for Actionable Intelligence to provide the mechanism for smart interactions for human-machine design in Industry and Services 5.0.

Since this is our first article related to intelligence, we want to be methodical in defining it and what it means to make it actionable. Intelligence requires ability to:

• sense the environment meaningfully,
• to make decisions, and
• to enable controlled action

Higher levels of intelligence will include the ability to recognize objects and events, to present knowledge in a world model with paradigms for world views, and to reason about the plan for the future.

In advanced forms, intelligence provides the capacity to perceive and understand the background and context for newly acquired facts or knowledge, to choose wisely, and to act successfully under a large variety of circumstances as to survive, prosper, and reproduce in a complex, sometimes hostile environment.

“We are all already cyborgs in a sense today. Phones and computers are extensions of ourselves…..”

The ultimate goal of Neuralink is to create a generalized I/0 interface for the brain. In the short term it is to help folks with debilitating issues. In the long term it is to create a benevolent interface with AGI by real time interactions with the brain. Our neural implants will be inconspicuous and enable intelligent interactions as we have seen with our studies with monkey pong and other animals” Elon Musk at Neuralink’s Show and Tell, Fall 2022.m“We have been on this very difficult journey from prototype to product. I’ve always said that protypes are easy, production is hard. It is one percent idea being easy and execution hard ”

The Intersection of biology and technology

When we talk about Actionable Intelligence at Numorpho Cybernetic Systems (NUMO), it is the combination of mechanistic activities with intelligent enactments that will be orchestrated by mind-machine interactions similar to what Neuralink is doing. We will also be investigating on using the Neuralink interface to bring another dimensionality to dynamic interactions for managing processes beyond and supplementing our current digital twine (digital threads + digital twins) approach.

In this article we will provide a background for AI and Cybernetics, define the essentials of what we call actionable intelligence by postulating a 5^th order of cybernetics.

BACKGROUND

The topic of artificial intelligence has been receiving a lot of attention recently for its development and application in many domains, growing in scope and complexity.

AI has been around for a long time, with the origins of AI tracing back to the Greek myth of Talos, the legendary giant bronze automaton who guarded the island of Crete. AI technology has been used in military warfare since WWI. The concept of AI is based on the human ability to reason, discover meaning, generalize, and learn from past experiences. AI is the result of applying cognitive science to create a computational model of human thought. AI technologies are used in a number of applications, including expert systems, natural language processing (NLP), speech recognition and machine vision. AI technology is now used in many smart applications that range from self-driving cars and drones to recommendation engines, search engines and social media.

Machine vision, which involves the ability of machines to recognize and understand visual information, is another important area of AI. This has enabled the development of self-driving cars, drones, and other systems that can navigate and interact with their environment using cameras and other sensors.

Overall, AI has the potential to revolutionize many different industries and has already made significant impacts in a variety of fields. As the technology continues to advance, it is likely that we will see even more widespread adoption and application of AI in the future.

The Constituents of AI

For the last fifty years, Artificial Intelligence (AI) has been one of the most fascinating fields in Computer Science. Unfortunately, there is a lack of understanding about what Artificial Intelligence really is. Partly this is because the field has been shrouded in mystery due to the absence of a rigorous mathematical theory – generally regarded by the AI community as an essential component for a proper scientific foundation. Also, partly because certain philosophical problems that are intimately related to AI, such as the nature of intelligence itself, and the relation between mind and body, remain unsolved.

AI is as useful as it is ubiquitous, but they are only as intelligent, rational, thoughtful, and unbiased as their creators. Safe, ethical, and effective AI will be vital for our future. This is an important issue in AI because systems are increasingly making decisions that have an impact on people’s lives, and we need to understand why they make the decisions they do.

An aspect that is becoming increasingly pertinent is about the use of AI both from training the neural nets and of utilizing the results. Perhaps there is need to understand their ramifications too in terms of its progression from Analytical, Actionable, Emotional and Existential on how it would interact with humans in terms of cooperation, collaboration, and even competition (coopetition?). Terms like responsible, ethical, fairness, explainable and interpretable need to be explored as use cases from an AI POV also. In each use case, both societal and technical aspects shape who might be affected by AI systems and how.

Cybernetics and its Evolution

Cybernetics is defined as the interaction between mind and machine. With more and more intelligent machines coming to the foray driven by AI, there is need to understand the ramifications this next industrial revolution brings to properly coordinate the complexities that arise in coordinating thought and action, intelligence, and the act of performing the task so that unintended circumstances are obviated.

Cybernetics is a wide-ranging field concerned with regulatory and purposive systems and has provided a general framework for analyzing communications and control processes in “purposeful” systems, from machines, to genomes to empires, to cognate beings. In our prior paper on Managing Complexity, we had outlined a framework to manage disparate information in unique schema structures to provide a blueprint for Cybernetic Mechanisms comprising of ingesting data, appropriately storing them, and utilizing a hybrid connective tissue to analyze and render them.

One of the key objectives in cybernetics is to seek a coherent theory for explaining the mechanisms of both natural and machine (artificial) intelligence and their correspondence. Intelligent Computing must be able to address novel situations and abstraction to automate ordinary human activities. Over the years (since 1950), cybernetics has evolved from:

• 1st order or Engineering Cybernetics
• 2nd order or Biological Cybernetics
• 3rd order or Social Cybernetics
• 4th order or Rational Cybernetics

Computational Cybernetics (CC) is the synergetic integration of Cybernetics and Computational Intelligence techniques. It covers not only mechanical, but biological, social, and economic systems, and for this uses intelligence-based results of communication theory, signal processing, information technology, control theory, the theory of adaptive systems, the theory of complex systems, and computer science. An example of how such integrated functioning in the human brain can be conceptualized in mathematical terms as a cybernetic system is detailed in the paper by Dr. S. S. Vinekar, MBBS published in 1968.

Artificial Intelligence (AI) and Cybernetics are widely misunderstood to be the same thing. However, they differ in many respects. For example, Artificial Intelligence (AI) grew from a desire to make computers smart, whether smart like humans or just smart in some other way. AI presumes that value lies in understanding “the world as it is” — which presumes that knowing the world is both possible and necessary. Cybernetics holds that it is only necessary to be coupled to the world sufficiently to gain feedback in order to correct actions to achieve a goal. Thus, while both fields must have clear and inter-consistent concepts such as representation, memory, reality, and epistemology (middle), there are more differences than similarities. One prominent difference is that Cybernetics is more concerned with control systems and feedback loops to maintain functions within certain range that may be considered optimum by the user. The actual knowledge of what is controlled is in the domain of AI. AI simulates the cognitive functions (higher cortical functions) while cybernetics simulates (subcortical physiological) regulatory functions (in the human body: such as blood pressure, pulse rate, body temperature, etc.) that take place below the cognitive levels of awareness. Cybernetics can be enriched with awareness, cognition, and knowledge but it is independent of these attributes.

The purpose for this treatise is to utilize the definition of Computational Cybernetics to:

• Enable the creation of increasingly complex but user-friendly cognitive structures that consist of qualitatively upgraded levels of integration with devices,
• Use software that can assimilate the perceptions and input received from the physical domain and instilled by intelligent human minds, with an ambition to extend the range of human perception far beyond imagination
• Coordinate their corresponding accurate representations and other operations in the digital domain, and process analog information therein, and
• Simulate the functioning of the human brain at the metaphysical level to help create a progressive order of intelligence for the next generation of cybernetic systems that are built with newer abilities for awareness.

These systems will then respond efficiently to stimulus (input) and provide timely tentative and when possible concrete pragmatic solutions.

Thus, Cybernetics offers a language (both vocabulary and frameworks) that enable scientists (and designers and others) from different domains of knowledge and practice to communicate—to describe the structural similarities of systems and to recognize patterns in information flows. This shared language is especially useful in analyzing, designing, and managing complex, adaptive systems, which are intertwined with many of today’s wicked problems.

In a subsequent article, we will detail out the Mantra M5^TM, a platform to manage cybernetic interactions to coordinate activities to make, manage, move, market and maintain – the five ms for process management. It will consist of a themed set of knowledge management artifacts that are dynamically rendered using Digital Twins and Augmented reality, Human-Machine Interfaces, and our forays in the Metaverse utilizing Web3.

The 5th Order of Cybernetics

We propose a 5th order for Contextual and Pragmatic Cybernetics which with be the basis for Existential Intelligence that will be discussed in subsequent articles. All of the definitions for Cybernetics provide for the needed impetus to act upon information and hence we define Cybernetics as Actionable Intelligence.

Defining Actionable Intelligence

Actionable intelligence is the result of applying cognitive science to create a computational model of human thought, and it is used to provide:

• A Decision Support System, which is a computer program that provides recommendations or decisions based on an analysis of data. The data may be from the organization or from the environment, and it may be structured or unstructured. The analysis may be simple or complex, and it may use a variety of data mining and machine learning algorithms.

and

• Machine intelligence – A mechanism for action, where robots with predefined articulations can perform the needed action safely, security and with rationality. When we talk about machine intelligence, we are referring to the ability of a machine to perform a task that requires human intelligence, such as reasoning, natural language understanding, and problem solving. Machine intelligence is a branch of AI that is concerned with the design and development of algorithms that enable computers to perform tasks that require human intelligence and appropriate articulations to perform the actions. Machine intelligence algorithms are used in a variety of applications, including expert systems, natural language processing, and computer vision.

In order to be actionable, intelligence has to be relevant and timely. The relevance of intelligence is determined by the user’s needs. The timeliness of intelligence is determined by the user’s timeframe. In order for intelligence to be actionable, it has to be both relevant and timely.

Thus:

Actionable intelligence is information that can be followed up on, with the further implication that a strategic plan should be undertaken to make positive use of the information gathered. Actionable intelligence must be relevant, timely, accurate, precise, and complete in order to be useful to the decision maker.

This definition will be explored through 4 concepts: AI, Actionable Intelligence, Actionable Intelligence Risk and AI Risk.

1. AI is the creation of tools through software, hardware or both that mimic human intelligence. AI can range from devices as simple as a calculator to much more complex tools as can be seen in today’s technology.
2. Actionable Intelligence is the capacity to understand the significance of AI, its progression, and decide on an action based on the information you have.
3. Actionable Intelligence Risk is the risk of taking or not taking action based on information you have. In 1950’s, computers were made that were capable of accomplishing mundane tasks but in 1980’s computers began to develop the capacity to exhibit human-level intelligence. While the IBM computer Deep Blue defeated world chess champion Garry Kasparov in 1997, it was not until 2011 that IBM’s supercomputer Watson beat two human Jeopardy! champions.
4. AI Risk – Now that we have some context, let’s explore 3 concepts to consider as AI Risk:
5. The first concept is bias. In 2015, the Pew Research Center conducted a study which found that 56% of respondents had a negative view of AI, 40% had a positive view, and 4% had a neutral view. Bias is a known issue in AI and is a major concern as tech becomes more engrained in society.
6. The second concept is “black box”. In 2015, Google’s AlphaGo defeated Lee Se-dol, the European Go champion in a best-of-5 series. AlphaGo was trained by Google using deep learning, an artificial intelligence technique that teaches computers to learn by example. Since that time, AlphaGo has been trained to play the Chinese strategy game called Wei chi (Exhibit C). AlphaGo’s success was attributed to deep learning, which Google described as an “unsupervised learning technique”. In supervised learning, humans must label data for the computer to learn, but in deep learning, the computers learn directly from the data itself. This ability to self-train makes the “black box” of AI a concern as it’s unclear how decisions are made.
7. The third concept is “explainability”. In 2016, IBM Watson Health partnered with Memorial Sloan Kettering Cancer Center to combine patient data with Watson’s learning capabilities. The goal was to help doctors understand and identify treatment options that were most appropriate for their patients. Initially, Watson was trained with 20,000 cancer-related articles and studies provided by Memorial Sloan Kettering. Watson was then asked to find all treatment options, assess each option’s potential success, and recommend a cancer treatment. However, soon after implementation, it was found that Watson’s results were often wrong, most significantly when recommending treatments for breast, lung, and colon cancers. After further examination, it was found that Watson’s results were highly dependent on the data available to it. For example, it treated skin cancers as lung cancers, treated bladder cancers as breast cancer.

In order to be actionable, intelligence has to be relevant and timely. The relevance of intelligence is determined by the user’s needs. The timeliness of intelligence is determined by the user’s timeframe. In order for intelligence to be actionable, it has to be both relevant and timely.

Thus, Actionable intelligence is information that can be followed up on, with the further implication that a strategic plan should be undertaken to make positive use of the information gathered. Actionable intelligence must be relevant, timely, accurate, precise, and complete in order to be useful to the decision maker.

Actionable intelligence is a process for making decisions. The basic question for Actionable intelligence is, “What do we need to know in order to make a decision?” In order to answer that question, we need to know what the decision is, what the options are, what the risks and rewards are, and what the impact of the decision will be.

• We need to have a model of the world that we can use to evaluate the options. The model of the world can be as simple as a mental model or as complex as a computer simulation. In order to have a model of the world, we need data. The data can be from sensors, from humans, or from other sources. This data has to be processed in order to be useful. The processing can be as simple as filtering the data or as complex as running a computer simulation. In order to process the data, we need algorithms. The algorithms can be simple or complex, and they can be static or dynamic.

• We need to have a In order to have a goal, we need to have a value system. The value system can be as simple as a utility function or as complex as a set of ethical principles.

• In order to decide, we need to have a way of choosing between the options. The choice can be as simple as a greedy algorithm or as complex as a genetic algorithm. In order to have a way of choosing between the options, we need to have a way of comparing the options. The comparison can be as simple as a linear function or as complex as a nonlinear function.

• The output of the decision-making process is the action. In order to take action, we need to have a way of translating the decision into an action. The process of actionable intelligence is iterative. The output of one step is the input to the next step. This process needs to be adaptive, flexible, scalable, and modular.

SUMMARY

Today, we stand at the cusp of a revolution that will change the way we think, do and survive in our world of tomorrow driven by technological innovations and the challenges that come with it. Albeit Industry 4.0 with its theme of smart manufacturing by including Internet of Things (IoT), Digital Thread, Digital Twin, AI, ML Big Data, Cloud Computing, and other technological advances has enabled rapid advances in making and managing products, there is still a long way to go to harmonize the value chain so that it is a seamless extension of our innate capabilities.

Imparting intelligence to machines in a way that it is not brittle (subject to a disastrous failure, when it happens), and enabling human-centric and resilient communication are needed for future cybernetic (man-machine) interactions. This goal for Industry and Services 5.0 will be the basis for NUMO’s actionable intelligence to enable rational, contextual, and pragmatic behaviors from machines, systems, and entities (previously described as earthlings, spacelings, and starlings). We will accordingly progress the theme for a 5th order of Cybernetics to achieve this state – a progression from the current AI techniques to what we call Existential Intelligence (EI).

NITIN UCHIL Founder, CEO & Technical Evangelist

nitin.uchil@numorpho.com