{
  "abstract": "Artificial intelligence research has primarily focused on constructing systems that perceive, reason, learn, plan, and solve problems. A deeper question concerns the origin and organization of these capabilities: can an intelligent system generate not only solutions, but also new mechanisms of intelligence, procedures for acquiring capabilities, and rules for constructing and adapting those procedures? This paper introduces Meta-Intelligence (MI) as a proposed general framework for investigating the generation and adaptation of mechanisms that produce intelligent behavior. The framework distinguishes task-level intelligence, capability construction, intelligence-mechanism construction, and the generation or modification of procedures governing these processes. It represents intelligent capabilities as executable or parameterized mechanisms that can be constructed, evaluated, selected, combined, and revised under explicit objectives and constraints. A formal model is proposed in which an intelligent system maintains a repertoire of capabilities and a meta-level controller that determines when existing capabilities are sufficient, when new mechanisms should be constructed, and when the procedures responsible for capability acquisition should be modified. The framework incorporates task performance, computational cost, uncertainty, robustness, and verification into a unified optimization objective. It also distinguishes generating new capabilities from merely selecting, prompting, or fine-tuning existing ones. Meta-Intelligence is positioned in relation to meta-learning, meta-reasoning, automated machine learning, cognitive architectures, and recursive self-improvement. These neighboring fields already address important aspects of the problem, including recent approaches that recursively improve problem-solving procedures. The proposed contribution is a general research framework for treating the construction of intelligence-producing mechanisms as an explicit, measurable computational objective. Testable hypotheses, an experimental methodology, and safety requirements are presented. No empirical superiority or emergence of general intelligence is assumed. The central research question is whether systems that can construct and evaluate new mechanisms of intelligent behavior can achieve more adaptable and resource-efficient performance than systems restricted to predefined capability repertoires.",
  "abstract_shown": true,
  "affiliations": [],
  "article_type": null,
  "arxiv": null,
  "author_total": 1,
  "authors": [
    {
      "name": "Jincheng Zhang",
      "orcid": "0009-0005-1860-0009"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/dhm75e772hg1e.bib",
  "cited_by": 0,
  "doi": "10.5281/zenodo.23498821",
  "fulltext_url": null,
  "id": "dhm75e772hg1e",
  "issns": [],
  "landing_url": "https://zenodo.org/records/23498821",
  "language": null,
  "licence": "CC-BY-4.0",
  "licence_label": "CC BY 4.0",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [],
  "openalex": null,
  "pdf_url": "https://zenodo.org/api/records/23498821/files/Meta-Intelligence%20A%20General%20Framework%20for%20Generating,%20Evaluating,%20and%20Adapting%20the%20Mechanisms%20of%20Intelligent%20Behavior.pdf/content",
  "pmcid": null,
  "pmid": null,
  "published": "2026-10-20",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/dhm75e772hg1e.ris",
  "sources": [
    {
      "first_seen": "2026-11-03T12:20:47Z",
      "native_id": "23498821",
      "source": "zenodo",
      "url": "https://zenodo.org/records/23498821"
    }
  ],
  "subjects": [],
  "title": "Meta-Intelligence: A General Framework for Generating, Evaluating, and Adapting the Mechanisms of Intelligent Behavior",
  "title_text": "Meta-Intelligence: A General Framework for Generating, Evaluating, and Adapting the Mechanisms of Intelligent Behavior",
  "type": "preprint",
  "updated_at": "2026-11-03T12:20:47Z",
  "url": "https://research.draconitic.com/paper/dhm75e772hg1e",
  "venue": "Zenodo",
  "versions": []
}