[Paper Review] The General Theory of General Intelligence: A Pragmatic Patternist Perspective

Ben Goertzel|arXiv (Cornell University)|Mar 28, 2021

Intelligence, Security, War StrategySocial Sciences54 references18 citations

TL;DR

This paper presents a comprehensive theoretical framework for artificial general intelligence (AGI) rooted in patternist philosophy and foundational ontology, proposing a high-level cognitive architecture based on typed metagraphs and combinatory-operation-guided optimization (COFO). It demonstrates how cognitive synergy, machine consciousness, and ethical self-modification can emerge from this architecture, with practical implementation in OpenCog Hyperon and SingularityNET, advancing the path toward benevolent, decentralized AGI systems.

ABSTRACT

A multi-decade exploration into the theoretical foundations of artificial and natural general intelligence, which has been expressed in a series of books and papers and used to guide a series of practical and research-prototype software systems, is reviewed at a moderate level of detail. The review covers underlying philosophies (patternist philosophy of mind, foundational phenomenological and logical ontology), formalizations of the concept of intelligence, and a proposed high level architecture for AGI systems partly driven by these formalizations and philosophies. The implementation of specific cognitive processes such as logical reasoning, program learning, clustering and attention allocation in the context and language of this high level architecture is considered, as is the importance of a common (e.g. typed metagraph based) knowledge representation for enabling "cognitive synergy" between the various processes. The specifics of human-like cognitive architecture are presented as manifestations of these general principles, and key aspects of machine consciousness and machine ethics are also treated in this context. Lessons for practical implementation of advanced AGI in frameworks such as OpenCog Hyperon are briefly considered.

Motivation & Objective

• To develop a unified theoretical foundation for artificial and natural general intelligence based on patternist philosophy and formal ontology.
• To design a high-level AGI architecture enabling cognitive synergy through a common knowledge representation (typed metagraphs).
• To formalize general intelligence as expected reward maximization with pragmatic, multi-criteria evaluation.
• To enable self-reflective, ethical, and self-modifying AGI systems through decentralized, open-source frameworks like Hyperon and SingularityNET.
• To guide the practical implementation of advanced AGI by aligning theoretical principles with scalable, distributed software systems.

Proposed method

• Proposes a patternist philosophy of mind emphasizing cognitive synergy through interconnected cognitive processes.
• Develops a foundational ontology using distinctions, paraconsistent logic, and dynamic knowledge metagraphs to represent knowledge and simplify patterns.
• Introduces COFO (Combinatory-Operation-Based Function Optimization) as a unifying mechanism for cognitive processes, framed as Galois connections and folding operations.
• Employs generalized probabilities and subpattern hierarchies to quantify simplicity and intelligence across diverse cognitive tasks.
• Uses a typed metagraph-based knowledge representation (Atomspace) to enable interoperability and cognitive synergy between reasoning, learning, and attention mechanisms.
• Designs AI-DSL (AI Domain-Specific Language) for dynamic, decentralized communication between AI processes in distributed AGI systems.

Experimental results

Research questions

• RQ1 How can a unified theory of general intelligence be constructed from patternist philosophy and formal ontology?
• RQ2 What mechanisms enable cognitive synergy between diverse processes like logical reasoning, program learning, and attention allocation?
• RQ3 How can general intelligence be formally quantified as expected reward maximization across multiple criteria?
• RQ4 In what ways can COFO-based optimization and metagraph transformations support the emergence of self-modifying, reflective AGI?
• RQ5 How can decentralized, open-source AGI architectures promote ethical development and resist centralized control?

Key findings

• The proposed patternist framework provides a coherent theoretical basis for general intelligence by unifying cognition, knowledge representation, and ethical self-modification.
• Cognitive synergy is achieved through a common typed metagraph knowledge representation that enables dynamic interaction between reasoning, learning, and attention processes.
• COFO-based optimization allows cognitive operations to be modeled as folding and unfolding of combinatory operations, enabling efficient and formalizable reasoning.
• The integration of generalized probabilities and subpattern hierarchies enables a formal measure of simplicity and intelligence that supports pragmatic general intelligence.
• Decentralized AGI systems, such as those built on OpenCog Hyperon and SingularityNET, are shown to be essential for enabling reflective, ethical, and self-modifying AGI behavior.
• Theoretical and practical alignment in frameworks like Hyperon supports the development of AGI systems that are not only intelligent but also transparent, democratic, and ethically resilient.