Category-Theoretic Comparative Framework for Artificial General Intelligence
This paper proposes a category-theoretic framework to compare different approaches to Artificial General Intelligence (AGI), aiming to formalize and relate diverse AGI architectures and theories through mathematical structures.
Background
- This paper is a formal, math-heavy preprint from arXiv (the open-access research repository for physics, math, and computer science). It uses **category theory** — a branch of abstract math that studies structures and their relationships via "objects" and "arrows" (morphisms) — to compare different approaches to AGI.
- **AGI (Artificial General Intelligence)** refers to a hypothetical future AI that can perform any intellectual task a human can, unlike today's "narrow" AI (e.g., ChatGPT, which excels at text but can't drive a car or cook). AGI is a major, controversial goal in AI research.
- The paper stays purely theoretical: it does not build a working AGI. Instead, it tries to create a mathematical language to define what "general intelligence" even means, and to compare competing architectures (e.g., deep learning vs. symbolic AI) on common ground. This matters because currently, different AI camps often talk past each other.