Article

Knowledge Graphs & General Collective Intelligence: Shifting to Industry 5.0

Posted August 10, 2022 | Technology | Amplify
KG_CGI
In this issue:

  AMPLIFY  VOL. 35, NO. 7
  
ABSTRACT

Andy E. Williams looks at how human-centric functional modeling (a way to allow computers to solve general problems) could be used to create KGs capable of providing compete semantic models of systems, enabling us to transition to Industry 5.0. He defines Industry 5.0 as a world in which far greater integration is possible, including functional computing approaches like GCI. Although the emergence of GCI isn’t guaranteed (it could end up in a technology gravity well), it would bridge type 1 and type 2 reasoning and lead to a radical increase in our ability to solve every problem.

 

The digital revolution has likely driven the single greatest transformation in the history of human civilization, but it might pale in comparison to the next great transformation: human-centric knowledge graphs (KGs) and functional computing approaches like general collective intelligence (GCI) that leverage such graphs.1

The digital revolution arose from the discovery that problems and solutions could be modeled digitally (i.e., in terms of ones and zeros) and from mass production of the transistors required to compute solutions to digital problems at exponentially greater speed and scale. But computer software or hardware created to solve one problem must be reengineered by people to solve another problem. Unlike the human brain, which has general problem-solving ability, computer hardware has only narrow problem-solving ability.

Human-centric functional modeling (HCFM) is a way to allow computers to solve general problems.2 HCFM represents problems via constructs called “functional state spaces.” These hypothetical functional state spaces are special types of KGs used to model systems. Functional state spaces are required for GCI and are of unprecedented importance if, as predicted, GCI can exponentially increase our capacity to understand systems.3 

For example, GCI has the potential to automatically reapply existing solutions to an exponentially greater number of different problems without any additional programming.4 Furthermore, the exponential increase in general problem-solving ability predicted to be possible through GCI applies to every process from design to recycling for every product or service that can be modeled with functional state spaces.

Taken a step further, GCI might be required to solve several key societal challenges corporations face. For example, humans can’t reliably discern social good — we tend to discern whatever matches the ideology to which our cognitive bias is predisposed.5 This may be why our current, non-GCI corporate environmental, social, and governance (ESG) programs have had limited success.6

Additionally, as technology advances, a phenomenon called the “technology gravity well” is expected to cause decision making to prioritize the interests of an ever-decreasing minority of individuals and businesses at the expense of achieving collective social good.7

It can take a long time for groups to understand when an individual is about to make (or has made) a decision that serves the group poorly (because of the sheer volume of decisions made by individuals within any large group and the often intractable number of potential interactions between those decisions). Looking out only for oneself is much easier, and those doing so might see benefit quickly.

Therefore, any force that acts to continually centralize decision making to an ever-decreasing minority of individuals and businesses would be expected to work to serve the interests of that minority far faster than collectively optimal choices can be understood and made by any group: a technology gravity well.

Some potential impacts of GCI (and therefore impacts of using HCFM to define KGs that are functional state spaces as required by GCI) relevant to this issue of Amplify on knowledge graphs are summarized in Table 1.

Table 1. The predicted impacts of using HCFM to define KGs that are functional state spaces  as required by GCI
Table 1. The predicted impacts of using HCFM to define KGs that are functional state spaces as required by GCI

Moving to Industry 5.0

Industry 4.0 refers to the transition to a world in which there is pervasive integration of manufacturing equipment and other physical systems with digital computing (cyber systems).

Industry 5.0 is defined here as the transition to a world in which far greater integration is possible through the use of HCFM to define KGs (functional state spaces) capable of providing complete semantic models of systems.

Assuming that human internal representations of anything that can be perceived with the physical senses as well as any emotion, thought, and conscious awareness can be defined in terms of functional state spaces through HCFM, then those spaces can be used to represent every physical object that can be seen and every idea that can be conceived. It follows, then, that HCFM potentially applies to every discipline from physics and mathematics to biology, psychology, computer science, and perhaps all others.8

In the design of products and services, the use of GCI implies the ability to explore a vastly larger region in the space of possible design configurations. This means exploring all possible permutations of all possible components and combinations of components. Rather than advancing through known research innovations, such design processes would mimic nature’s process of designing living things: incorporating any change that increases the likelihood of achieving an objective, even when the mechanisms by which that increase was achieved remain unknown.

In manufacturing, modeling physical products in terms of functional state spaces implies the ability to accommodate manufacturing processes too complex to be understood by any individual process designer in an effort to achieve competitive advantage.

In recycling, modeling products and services in terms of functional state spaces and the use of GCI might enable sustainability solutions that are impossible otherwise, such as radically reducing greenhouse gases through an exponential reduction in consumption. This could result from an ecosystem of GCI-based products that cooperate to become more durable, reusable, and recyclable than could be accommodated by any business model today.

A Peek into the Future

It might seem like HCFM and GCI could be used to improve any product or service. However, research suggests that managers can’t lead teams effectively when those managers are too smart (i.e., an IQ that is 1.2 standard deviations or higher than that of the group).9

So what might be the consequence of deploying a GCI with an IQ billions of points greater than that of the smartest human who has ever lived?10 One possible outcome is a technology design process in which human contributions come together too quickly and in ways too complex for any human to understand, resulting in technology so complex it can’t be reliably distinguishable from magic.

GCI-based technology will likely also be different in how it’s used. Rather than having to learn to operate such technology, the technology might learn what we are trying to do and self-assemble from available components to accomplish our goals to the optimal degree, removing the need for any specialized tools or expertise.

Although hard to envisage, this implies GCI might make individuals who are novices much more effective at specialist tasks like product design than the most gifted designers today and allow people with no medical training to perform surgeries and other medical interventions that the most gifted of today’s doctors would consider miraculous.

Just like the revolutionary digital technology that came before it, through simple geometric arguments in conceptual space, the case can be made that GCI is the most important technological development in the history of human civilization with regard to problems that can be modeled in terms of functional state spaces, which potentially includes all problems.

GCI might be as profoundly important as this (seemingly preposterous) claim, or it might prove impossible. However, observation of natural systems such as our own human organism suggests that adaptive problem-solving systems based on functional state spaces such as GCI are a real possibility. Nature has already created such solutions, and they have proven successful for hundreds of millions of years.

Escaping the Technology Gravity Well

The emergence of GCI isn’t a certainty. An analysis based on HCFM predicts that any civilization might go one of two ways. The first is to develop a mechanism for individual optimization, a necessarily centralized process that eventually might exponentially increase our ability to solve problems for companies or other individual entities. The second is to develop a mechanism for collective optimization, a necessarily decentralized and distributed process that eventually might exponentially increase the ability to solve problems for all.

The first option implies a civilization that will fall deeper into the technology gravity well toward the emergence of AGI, which is predicted to act as an exponentially more powerful system of individual optimization that makes a system of collective optimization like GCI impossible.11

Since this fall into the technology gravity well is likely to be accompanied by the removal of protections against abuse while radically increasing the ability for corporations, governments, and other entities to be abusive, this suggests unprecedented levels of abuse and control on the part of the company that falls to the bottom of the well first.

This would mean a negative outcome for every business except the one at the top of the hierarchy, which would be expected to gain all possible technological advantages to control more revenue than any company that has ever existed.

The other option would be using GCI to escape the technology gravity well, resulting in a positive outcome for the majority of businesses (except those that decide to fight this transition rather than embrace the far larger opportunities expected to come with it). Because GCI creates potentially unbeatable competitive advantage, companies that fight GCI would most likely go extinct.

Functional state space is involved in both transitions (to AGI or GCI). Even though functional state spaces and GCI have not yet been fully implemented, if it’s true that they have the greatest potential for impact on all technologies known today, then it’s important they are on every business leader’s radar.

Unlike today’s databases, which can only store a limited subset of information, functional state space has the potential to model any system and all possible behaviors of that system, potentially storing all possible information about a given system. Thus, a functional state space is a complete semantic model that enables meaning (understanding) rather than just information to be communicated at exponentially greater speed and scale.

But even without any understanding of functional state space, it is possible to use patterns of collectively intelligent cooperation to reliably achieve a radical increase in the ability to solve any problem.12 These patterns leverage a set of well-defined and generalizable relationships between businesses, their products or services, and other entities, without the need to recognize these relationships as existing in functional state space at all (though representing those relationships this way might allow them to be further generalized to achieve more impact).

The most stunning claim of GCI is that for certain categories of “wicked” problems (like achieving social good in difficult cases), the more we’re fixated on solving these problems, the less we are able to do so. Problems too complex to be solved directly through any choices that can be deduced by any individual must be solved indirectly through development of a more powerful distributed problem-solving system (such as GCI) that is capable of discovering far more complex choices that might be capable of radically better outcomes. All problem-solving methods that are not orchestrated by GCI can be considered direct (in that choices are generated by individuals) and centralized (in that these individuals can’t be prevented from prioritizing their own interests). This is problematic because no direct approach can reliably solve wicked problems (those currently considered not solvable).13 This is supported by the fact that no approach has reliably created durable solutions to these problems at any time in the history of human civilization. That means focusing energy toward any efforts other than GCI is the best way to not solve the world’s most challenging problems.

This is counterintuitive, since it would mean people who want to solve complex problems of social good might be the ones ensuring that the most pressing problems of social good cannot be solved. That is, when these individuals believe they know the solution, they don’t ensure that they or someone else diligently explores whether or not it is feasible to achieve an exponential increase in impact on social good through modeling problems and solutions in terms of functional state spaces/knowledge graphs together with the use of GCI. In other words, their caring ensures those problems can’t be solved.

Opposition from such well-meaning individuals is thus an important consideration when attempting to launch any GCI-based initiative if it is true that people interested in social good are predisposed to having cognitive biases toward type 1 reasoning (intuitive), preventing them from asking whether or not disruptive new solutions like GCI are needed.

Similarly, the institutions we rely on for coordinating social good globally use intuitive reasoning, making it impossible for them to choose interventions like GCI that are not similar to patterns of interventions in the past. Thus, any plan to achieve a radical impact on social good must consider working outside such institutions.

These innate cognitive biases are also an important factor to consider if, in addition, the type 2 reasoning (rational) that allows individuals to assess radically different solutions is typically not effective at building the mindshare required to build the consortia and attract the resources necessary to implement such an idea.

Any implementation of GCI might bridge these two reasoning types when it becomes available, but implementing GCI is the precise problem we’re trying to solve. The only solution might lie in understanding how nature has evolved complex adaptive systems in an iterative way, so that GCI could be implemented incrementally to bridge these reasoning types while enabling the implementation of a larger subset of GCI functionality.

GCI in turn requires implementing KGs that meet the requirements of functional state spaces, a problem that hasn’t yet been solved. By informing a variety of stakeholders (especially mathematicians, physicists, computer scientists, and others who study constructs with similar features) about how the combination of GCI and functional state spaces might radically increase our ability to solve every problem in general, it might be possible to inspire a collaborative effort to solve this problem as well.

References

Williams, Andy E. “Defining a Continuum From Individual, to Swarm, to Collective Intelligence, and to General Collective Intelligence.” International Journal of Collaborative Intelligence, Vol. 2, No. 3, 2 May 2022.

Williams, Andy E. “Automating the Process of Generalization.” AfricArXiv, 12 March 2022.

Williams, Andy E. “Human-Centric Functional Modeling and the Unification of Systems Thinking Approaches: A Short Communication.” Journal of Systems Thinking, 20 August 2021.

Williams, Andy E. “Cognitive Computing and Its Relationship to Computing Methods and Advanced Computing from a Human-Centric Functional Modeling Perspective.” SCRS Conference Proceedings on Intelligent Systems. SCRS Publications, 21 September 2021.

Williams, Andy E. “Innate Collective Intelligence and the Collective Social Brain Hypothesis.” PsyArXiv, 26 May 2022.

Williams, Andy E. “General Collective Intelligence as a Platform for Computational Social Systems.” AfricArXiv, 12 March 2022.

Williams, Andy E. “Are Wicked Problems a Lack of General Collective Intelligence?AI & Society: Journal of Knowledge, Culture, and Communication, 4 October 2021.

Williams, Andy E. “Human-Centric Functional Modeling and the Metaverse.” Journal of Metaverse, Vol. 2, No. 1, 29 April 2022.

Antonakis, J., R.J. House, and D.K. Simonton. “Can Super Smart Leaders Suffer from Too Much of a Good Thing? The Curvilinear Effect of Intelligence on Perceived Leadership Behavior.” Journal of Applied Psychology, Vol. 102, No. 7, 2017.

10 Williams, Andy E. “Defining and Quantifying an Exponential Increase in General Problem-Solving Ability Within Groups.” AfricArXiv, 22 February 2022.

11 Williams, Andy E. “Breaking Through the Barriers Between Centralized Collective Intelligence and Decentralized General Collective Intelligence to Achieve Transformative Social Impact.” International Journal of Society Systems Science, forthcoming 2022.

12 Williams, Andy E. “Increasing the Societal Impact of Science, Technology, Engineering, and Math with General Collective Intelligence.” AfricArXiv, 2 March 2022.

13 Williams (see 7).

 

About The Author
Andy E Williams
Andy E. Williams is chairman and CTO of Nobeah Technologies and founder and Executive Director of Nobeah Technologies Foundation. He is an expert in general collective intelligence (GCI) and human-centric functional modeling (HCFM). As a social entrepreneur, Mr. Williams focuses on understanding the equations underlying human challenges, so that by changing those equations, the problems solve themselves. His research on how to overcome the… Read More