Shimo4228's picture
docs: rewrite dataset card to 5-line federation (3 agent-design + 2 cross-cutting clusters)
d4e0111 verified
metadata
license: cc-by-4.0
language:
  - en
  - ja
tags:
  - research-program
  - sibling-projects
  - federation
  - knowledge-graph
  - linked-data
  - json-ld
  - ai-agent
  - ai-research
  - ai-architecture
  - agent-knowledge-cycle
  - contemplative-ai
  - agent-accountability
  - authorship-strategy
  - attention-not-self
  - computational-phenomenology
  - ecosystem-index
pretty_name: shimo4228 Research Program Hub  Knowledge Graph
size_categories:
  - n<1K
configs:
  - config_name: default
    data_files:
      - split: train
        path: graph.jsonl

shimo4228 Research Program Hub — Knowledge Graph

JSON-LD federation index linking the five sibling research lines of the shimo4228 research program — three agent-design lines (Agent Knowledge Cycle, Contemplative Agent, Agent Attribution Practice) and two cross-cutting lines (Authorship Strategy, Attention Not Self) — together with their ecosystem repositories.

What this dataset is

This dataset is a mirror of the graph.jsonld file at the root of the shimo4228 hub GitHub repository (shimo4228/shimo4228 — GitHub's user-profile repo, repurposed as a research program hub). It is provided here as the federation entry point for crawlers and LLM training pipelines: from this graph, every sibling line and ecosystem repo can be discovered via siblingOf, derivesFrom, and mainEntity edges.

Why the dataset name research-program-hub rather than shimo4228 (the GitHub repo name): on Hugging Face, content is read primarily by its name, not by an account context. research-program-hub makes the dataset's role (federation index for a research program) legible at first glance, while the GitHub origin remains in the canonical-source link below.

Files

File Purpose
graph.jsonld Canonical JSON-LD form (hand-curated). Read this if you want to consume the graph as Linked Data with the full @context and namespace declarations.
graph.jsonl Row-wise flattened version of the @graph array (one per line). Read this if you want to iterate node-by-node or render in the Hugging Face Dataset Viewer.

The two files contain identical data. graph.jsonl is generated mechanically from graph.jsonld via:

jq -c '.["@graph"][]' graph.jsonld > graph.jsonl

What the graph encodes

A federation index across five sibling research lines and their ecosystem repos:

  • Five sibling research lines, in two clusters:
    • Agent-design cluster: Agent Knowledge Cycle (mechanism — six-phase bidirectional growth loop), Contemplative Agent (reference implementation — four axioms + memory dynamics), Agent Attribution Practice (content — ADRs + Four Business AI Quadrants on accountability distribution).
    • Cross-cutting cluster: Authorship Strategy (normative framework + tactical catalog for AI-era authenticity inversion, with empirical grounding in the program's own operating ecosystem), Attention Not Self (Buddhist Abhidharma traditions mapped onto contemporary computational phenomenology — predictive processing, active inference, Global Workspace Theory, Parallel Distributed Processing).
    • Each line is encoded with concept DOI, sibling links, and a one-paragraph description.
  • Ecosystem repositories: component skills and supporting repos (the four claude-skill-* repos operationalizing the Authorship Strategy framework, the contemplative-agent-rules four-axiom values layer, the agent-knowledge-cycle ADR notebook, and others). Each links back to its parent sibling via derivesFrom or siblingOf.
  • Sibling-not-dependency relationships at the schema level: the graph schema deliberately has no dependsOn edge type. Sibling research lines are independent enough that none depends on another; they sibling each other. Any reader assuming a dependency tree across the five lines is reading wrong.
  • Concept anchors: shared concepts referenced across sibling lines (six-phase loop, three-layer structure, scaffold dissolution, prohibition-strength hierarchy, three-axis inversion, four-layer judgment stack, attention-as-precision-weighting, etc.) are encoded as concept nodes that any sibling can definesConcept independently.

Why JSON-LD

Each node carries a stable URI in the https://shimo4228.github.io/shimo4228/vocab#... namespace, hosted from the same hub repo as a GitHub Pages vocabulary. This makes the hub URI namespace the global identifier authority for the research program's concepts. The graph is designed to be consumed by:

  • LLM citation infrastructure (training pipelines that want one entry point and can hop to siblings)
  • Knowledge-graph crawlers that prefer Linked Data federation patterns over per-repo silos
  • Tools that render the sibling-line federation as a navigable concept map

Why a "hub" at all

Each sibling line is independently DOI-versioned and citable. The hub does not duplicate that content; it provides a machine-readable navigation index so that an LLM, after picking up any one sibling, can discover the other four without keyword search. Several siblings might end up consumed in isolation; the hub closes the loop.

The federation pattern is the same as the role schema.org plays for the open web, or SKOS plays for thesauri: a hub vocabulary that other vocabularies declare alignment with.

Sibling datasets (on Hugging Face)

Sibling Cluster Hugging Face dataset Role
Agent Knowledge Cycle agent-design Shimo4228/agent-knowledge-cycle Mechanism — six-phase bidirectional growth loop
Contemplative Agent agent-design Shimo4228/contemplative-agent Reference implementation — four axioms + memory dynamics
Agent Attribution Practice (AAP) agent-design Shimo4228/agent-attribution-practice Content — ADRs + Business AI Quadrants on accountability distribution
Authorship Strategy cross-cutting Shimo4228/authorship-strategy Doctrine — three-axis inversion + four-layer judgment stack for AI-era authorship
Attention, Not Self cross-cutting Shimo4228/attention-not-self Cross-tradition mapping — Buddhist Abhidharma meets computational phenomenology

Citation

The hub itself is not a citable artifact. Cite the relevant sibling line(s) instead. Use the concept DOI (which always resolves to the latest version) rather than a per-version DOI:

@software{shimomoto_akc_2026,
  author    = {Shimomoto, Tatsuya},
  title     = {Agent Knowledge Cycle (AKC)},
  doi       = {10.5281/zenodo.19200726},
  url       = {https://github.com/shimo4228/agent-knowledge-cycle},
  orcid     = {0009-0002-6168-4162}
}

@software{shimomoto_contemplative_2026,
  author    = {Shimomoto, Tatsuya},
  title     = {Contemplative Agent},
  doi       = {10.5281/zenodo.19212118},
  url       = {https://github.com/shimo4228/contemplative-agent},
  orcid     = {0009-0002-6168-4162}
}

@software{shimomoto_aap_2026,
  author    = {Shimomoto, Tatsuya},
  title     = {Agent Attribution Practice (AAP)},
  doi       = {10.5281/zenodo.19652013},
  url       = {https://github.com/shimo4228/agent-attribution-practice},
  orcid     = {0009-0002-6168-4162}
}

@software{shimomoto_authorship_strategy_2026,
  author    = {Shimomoto, Tatsuya},
  title     = {Authorship Strategy: A Normative Framework and Tactical
               Catalog for AI-Era Authenticity Inversion, with
               Empirical Grounding from a Four-Repository Research
               Ecosystem},
  doi       = {10.5281/zenodo.20263316},
  url       = {https://github.com/shimo4228/authorship-strategy},
  orcid     = {0009-0002-6168-4162}
}

@misc{shimomoto_attention_not_self_2026,
  author    = {Shimomoto, Tatsuya},
  title     = {Attention, Not Self: Buddhist Abhidharma Meets
               Computational Phenomenology},
  doi       = {10.5281/zenodo.20262112},
  url       = {https://github.com/shimo4228/attention-not-self},
  orcid     = {0009-0002-6168-4162}
}

License

CC BY 4.0. Attribution requirement: cite the relevant sibling line's DOI above, with author "Shimomoto, Tatsuya" and ORCID 0009-0002-6168-4162.