The Need for General Collective Intelligence in Addressing the Sustainable Development Goals

2020

Preprint

Self Cite

Collective intelligence algorithms aim to harness the intelligence of crowds in selecting optimal solutions to specific problem. General collective intelligence has been defined in a Functional Modeling Framework for representing and comparing models of individual and collective cognition as collective intelligence with general problem solving ability. A theoretical model of General collective intelligence is developed in this paper according to that framework. For each element of functionality required by cognition as defined in the Functional Modeling Framework, this paper defines a boundary condition that any potential solution must satisfy, defines a solutions that potentially satisfies all of these boundary conditions, and then defines criteria for that solution to be applicable to a system of collective cognition (a General Collective Intelligence or GCI). Where a collective intelligence algorithm might leverage the intelligence of crowds to select optimal solutions, this model of General Collective Intelligence aims to tackle the problem of not only selecting the optimal solution, including the optimal collective intelligence algorithm in each problem domain, but also selecting the optimal problem to solve.

Section: Validation Of the Model And Proposed Implementation In Softwmentioning

confidence: 99%

A Model for General Collective Intelligence

2020

Preprint

Self Cite

“…Rather than maximizing collective outcomes, allowing processes to become centralized forces decision-making to align with the interests of a subset of decision-makers, often working against maximizing collective outcomes [11]. Secondly, while an AI maximizes impact on a problem, the adaptation processes defined by this model are proposed to be required to enable an AGI to maximize outcomes by choosing the optimal problem to solve [15].…”

Section: The Validity Of a Functional Modeling Approach Towards Reprementioning

confidence: 99%

“…General problem solving ability from this perspective becomes a sequence of reasoning through which the mind can sustainably navigate the conceptual space in a way that potentially enables it to navigate from any concept to any other concept, that is, which gives the mind the ability to potentially solve any problem it can conceive of. In the model of AGI defined within the FMF, this stability is achieved by using the Lorentz equations to find the optimal problem to solve, so this library of problem solving functions can potentially be used to gain general problem solving ability [10], [14], [15].…”

Section: General Problem Solving Ability In the Functional Modeling Fmentioning

confidence: 99%

Defining Functional Models of Artificial Intelligence Solutions to Create a Library that an Artificial General Intelligence can use to Increase General Problem Solving Ability

2020

Preprint

Self Cite

The AI industry continues to enjoy robust growth. With the growing number of AI algorithms, the question becomes how to leverage all these models intelligently in a way that reliably converges on AGI. One approach is to gather all these models ingo a single library that a system of artificial intelligence might use to increase it's general problem solving ability. This paper explores the requirements for building such a library, the requirements for that library to be searchable for AI algorithms that might have the capacity to significantly increase impact on any given problem, and the requirements for the use of that library to reliably converge on AGI. This paper also explores the importance to such an effort of defining a common set of semantic functional building blocks that AI models can be represented in terms of. In particular, how that functional decomposition might be used to organize large scale cooperation to create such an AI library, where that cooperation has not yet proved possible otherwise. And how such collaboration, as well as how such a library, might significantly increase the impact of each AI and AGI researcher’s work.

“…In short, GCI is essentially a system for maximizing collective outcomes, but since by design a GCI is required to be a dynamically stable system with respect to available resources, and since any outcome in general might be required to achieve or restore stability, GCI must must also maximize the degree to which it can achieve collective outcomes per unit of resources. According to the model of GCI, without such a system all current decision systems, whether autocratic rule or democratic voting, contain features or bugs that align decision-making with the interests of some subset of individuals rather than with the collective well-being (Williams, 2020a), (Williams, 2020e). This holds both for decision-making on the political left in the US, as well as on the political right, and is especially apparent during times of crisis that draw attention to flaws in decision-making, such as the current COVID-19 pandemic (Williams, 2020g).…”

Section: Defining General Collective Intelligencementioning

confidence: 99%

Bridging the Left - Right Political Divide with Artificial Intelligence and Collective Intelligence

2020

Preprint

Self Cite

A model of cognition suggests that the left vs right political debate is unsolvable. However the same model also suggests that a form of collective cognition (General Collective Intelligence or GCI) can allow education, health care, or other government services to be customized to the individual, so that individuals can choose services anywhere along the spectrum from socialized services if they desire, or private services if they desire, thereby removing any political stalemate where it might prevent any progress. Whatever services groups of individuals choose, GCI can significantly increase the quality of outcomes achievable through either socialized or private services today, in part through using information regarding the fitness of any services deployed, to improve the fitness of all services that might be deployed. The emerging field of General Collective Intelligence (GCI) explores how platforms might increase the general problem-solving ability (intelligence) of groups so that it is significantly higher than that of any individual. Where Collective Intelligence (CI) must find the optimal solution to a problem or group of problems, having general problem-solving ability, a GCI must also have the capacity to find the optimal problem to solve. In the case of political discussions, GCI must have the ability to re-frame political discourse from being focused on questions that have not proved resolvable, such as whether or not left leaning or right leaning political opinions are in general more “right” or “wrong”. Instead GCI must have the ability to refocus discussions, including on how to objectively determine whether a left or right bias optimizes outcomes in a specific context, and why. This paper explores the conjecture that determining whether a left leaning or right leaning cognitive bias is "optimal" (i.e. "true) based on any CI or other aggregate of individual reasoning that is not GCI, cannot reliably converge on "truth" because each individual cognitive bias leads to evaluating truth according to different reasoning types (type 1 or type 2) that might give conflicting answers to the same problem. However, through using functional modeling to create the capacity to represent all possible reasoning processes, and through using functional modeling to represent the domains in conceptual space in which each reasoning process is optimal, it is possible to systematically categorize an unlimited number of collective reasoning processes and the contexts in which execution of those reasoning processes with a right leaning or left leaning bias is optimal for the group. By designing GCI algorithms to incorporate each bias in its optimal context, a GCI can allow individuals to participate in collective reasoning despite their biases, while collective reasoning might still converge on "truth" in terms of functioning to optimize collective outcomes. And by deploying intelligent agents incorporating some subset of AGI to interact on the individual's behalf at significantly higher speed and scale, collective reasoning might gain the capacity to consider all reasoning and all "facts" available to any individual in the group, in order to converge on that truth while significantly increasing outcomes.