Computational Care Ecology: Synthetic Experiments and Internal Verification of CCE-MVM v0.1

Daiki Morimoto

Abstract

Background: Computational Care Ecology (CCE), defined in this study, separates information, proposals, responses, decision rights, and outcomes. Objective: To specify and internally verify CCE-MVM v0.1. Methods: No human data were used; inputs were synthetic. It has person, AI, family, and professional components. M0 records one decision event with observations, person models, proposals, an exogenous person response, six CCE rights gates, and actor-specific outcomes; M1 examines person model updating; M2 examines synthetic AI proposal-response coupling under rights conditions. Analytical verification targets were H1a, H2a, and H1b-1; their synthetic runs were implementation checks, not independent experimental evidence for those algebraic consequences. H2b was a recovery check limited to the stated conditions; H1b-2 was a scoped check of the decision rule and rights cases. Results: The H1a and H2a implementation checks reproduced more persistent error at lower learning rates after permanent change and greater tracking of temporary observations at higher rates. The H1b-1 implementation check matched the specified coupling-response relation. H2b showed faster recovery ordering at higher rates only for shock length 2 and tolerance 0.10; recovery was not measured under noise. The scoped H1b-2 checks matched the specified outcomes for response, option, and six rights gates. In author-reported internal verification, 96 core model tests passed; 41 future extension tests brought the development suite to 137 tests; CSV and JSON outputs matched during internal regeneration. Conclusions: Author-reported internal verification found CCE-MVM v0.1 internally consistent with its Specification under the tested conditions. This is verification, not empirical, clinical, ethical, or safety validation. It does not model endogenous interaction among four actors; external validity remains unestablished.

全文PDFを読む

原稿の日付は2026年9月16日。未査読の研究原稿であり、学術誌やarXivでの掲載を示すものではない。日本語記事の第6章までで説明する三つのモデルの基礎を扱い、将来構想全体を検証したものではない。

引用情報

Morimoto D. Computational Care Ecology: Synthetic Experiments and Internal Verification of CCE-MVM v0.1. Unknown Lab; 2026. Version 2026-09-16. https://unknownlab.pages.dev/publications/cce-mvm-v0-1/

文書種別
英語研究原稿・未査読
DOI
未付与
ORCID
0009-0009-6444-750X
ライセンス
本文の権利表記を参照

関連するサイト記事を読む · 研究資料一覧