This page is the public publication registry for S. Meta research items released on Jxiv, JST Preprint Server.
このページは、S. Meta の研究のうち Jxiv, JSTプレプリントサーバ で公開された項目をまとめた公開レジストリです。
Status note / 状態注記: These are Jxiv preprints or public research artifacts. Jxiv publication does not by itself mean peer review or institutional validation. / これらはJxiv上のプレプリントまたは公開研究成果です。Jxivでの公開それ自体は、査読済みであることや制度的な妥当性認定を意味しません。
Japanese title: 生成AI出力における説明整合性と現実接触の乖離を点検する補助枠
Author: S. Meta
Jxiv DOI: 10.51094/jxiv.4696
Published: 2026-07-06
Role in the archive: A cross-cutting audit concept for distinguishing explanatory coherence from contact with reality.
Archive pages:
Primary distinction:
Coherence is not contact.
Japanese title: 構造変化センサー（Structure-Change Sensor）：時系列対話・文書における構造変化候補の検出とリアルタイム・ブックマーキング
Author: S. Meta
Jxiv DOI: 10.51094/jxiv.6215
Submitted: 2026-08-26
Published: 2026-09-08
Role in the archive: A method for detecting and preserving candidate structural-change points in longitudinal dialogue, documents, and interaction records without converting detection directly into interpretation or causation.
Archive page:
The public method distinguishes:
Detection ≠ Interpretation ≠ Transmission ≠ Causation
Candidate ≠ verdict
The retrospective sensor is supported diagnostically across multiple cases; the realtime extension is specified and in prospective trial use rather than presented as a validated measurement scale.
Scope note: The paper is not limited to human-AI dialogue. Its public formulation covers longitudinal dialogue, documents, and interaction records, while human-AI collaboration is one of its explicit research contexts.
Japanese title: 実証監査アーキテクチャ（EAA）：複数AIによる証拠再構成を比較可能にする監査方法の開発と前向き検証
English title: Empirical Audit Architecture (EAA): Development and Prospective Validation of an Audit Method for Comparable Evidence Reconstruction by Multiple AI Systems
Author: S. Meta
Jxiv DOI: 10.51094/jxiv.6148
Published: 2026-09-16
Role in the archive: An audit method for comparing evidence reconstruction by multiple AI systems while preserving original outputs, differences, uncertainty, and replayable audit conditions.
Archive page:
EAA was developed through development cases and then applied to a new case after freezing the method specification. In the paper, “prospective validation” refers to application to a new case after method freeze; it does not imply full preregistration of the entire case-selection process.
Primary distinction: comparable reconstruction requires more than similar conclusions. Evidence scope, uncertainty, normalization, and execution conditions must remain inspectable.
EAA and SC are distinct but complementary public methods.
They should not be collapsed into a single framework, and publication of one does not validate the other.
Within the broader archive, they sit near the AAS / Structural Drift research track because both concern auditability across changing records, interpretations, and long-horizon human-AI work. They do not replace AAS, Reality Stabilizer, or the core concept map.
For archive-level relationships, motivation, and scope boundaries, read:
The publication registry is an index of released research artifacts. It is not a substitute for the individual papers, their supplementary materials, or their stated limitations.