Beyond Self-Diagnosis and Mutual Diagnosis: A Universal Adaptive Self-Mutual-Human Diagnostic Framework for Dynamic Error Detection, Assessment, and Switching Among Artificial Intelligence Models and Human Participants

Existing research on artificial intelligence diagnosis, self-evaluation, error detection, model criticism, multi-agent assessment, and human oversight has investigated different approaches to identifying errors, weaknesses, inconsistencies, and potential failures in artificial intelligence systems. However, these approaches are frequently examined through separate operational configurations, such as a model inspecting its own outputs, multiple models evaluating one another, or humans reviewing artificial intelligence-generated results. A general theoretical framework that unifies these three diagnostic dimensions while enabling context-dependent selection, dynamic switching, simultaneous combination, and structural reconfiguration remains an important research direction. This paper introduces the Universal Adaptive Self-Mutual-Human Diagnosis Framework (UASMH-DF), a mechanism-independent diagnostic architecture grounded in the Universal Self-Mutual-Human Model (USMHM). The framework establishes three fundamental diagnostic dimensions: Self-Diagnosis, in which an individual artificial intelligence model or participating entity identifies and assesses problems within its own outputs, processes, decisions, or internal operational results; Mutual Diagnosis, in which two or more independent participants examine, evaluate, and identify potential problems in one another's outputs or activities; and Human Diagnosis, in which human individuals or groups participate in identifying, examining, interpreting, and assessing potential problems involving artificial intelligence systems or other participants. The framework does not assume that any single diagnostic dimension is universally sufficient or optimal. Instead, it introduces Universal Adaptive Diagnostic Selection and Switching, enabling an individual system or coordinating architecture to determine whether a diagnostic task should be conducted through self-based, mutual-based, human-based, or hybrid configurations. The diagnostic configuration can be dynamically modified according to task complexity, observed diagnostic performance, uncertainty, error severity, participant capabilities, disagreement, available evidence, resource constraints, and governance requirements. Furthermore, the framework distinguishes diagnosis from correction, diagnostic capability from diagnostic reliability, adaptation from switching, and simultaneous diagnostic combination from structural reconfiguration. It introduces a formal representation of diagnostic processes, a three-dimensional diagnostic configuration space, adaptive decision policies, dynamic transition mechanisms, diagnostic lifecycle, evaluation principles, and governance requirements. The framework supports a broad range of diagnostic functions, including factual error detection, logical inconsistency identification, reasoning assessment, hallucination detection, instruction-compliance diagnosis, security assessment, bias identification, procedural verification, decision evaluation, and system-performance diagnosis. To the best of our knowledge, this paper is the first to formally propose a unified, mechanism-independent theoretical framework that integrates Self-Diagnosis, Mutual Diagnosis, and Human Diagnosis with universal adaptive diagnostic selection, dynamic switching, combination, and reconfiguration within the broader USMHM architecture. Rather than asserting that the individual diagnostic mechanisms are unprecedented, this priority claim concerns their proposed unified organization and adaptive operational architecture. The framework provides a general conceptual foundation for designing and evaluating dynamically adaptive diagnostic systems involving heterogeneous artificial intelligence models and human participants.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-18
DOI
https://doi.org/10.5281/zenodo.24337743
Primary Topic
Explainable Artificial Intelligence (XAI)
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Beyond Self-Diagnosis and Mutual Diagnosis: A Universal Adaptive Self-Mutual-Human Diagnostic Framework for Dynamic Error Detection, Assessment, and Switching Among Artificial Intelligence Models and Human Participants

Jincheng Zhang
4 links
Zenodo (CERN European Organization for Nuclear Research)
Explainable Artificial Intelligence (XAI)
preprint

Beyond Self-Diagnosis and Mutual Diagnosis: A Universal Adaptive Self-Mutual-Human Diagnostic Framework for Dynamic Error Detection, Assessment, and Switching Among Artificial Intelligence Models and Human Participants

Jincheng Zhang
preprint en

Abstract

Existing research on artificial intelligence diagnosis, self-evaluation, error detection, model criticism, multi-agent assessment, and human oversight has investigated different approaches to identifying errors, weaknesses, inconsistencies, and potential failures in artificial intelligence systems. However, these approaches are frequently examined through separate operational configurations, such as a model inspecting its own outputs, multiple models evaluating one another, or humans reviewing artificial intelligence-generated results. A general theoretical framework that unifies these three diagnostic dimensions while enabling context-dependent selection, dynamic switching, simultaneous combination, and structural reconfiguration remains an important research direction. This paper introduces the Universal Adaptive Self-Mutual-Human Diagnosis Framework (UASMH-DF), a mechanism-independent diagnostic architecture grounded in the Universal Self-Mutual-Human Model (USMHM). The framework establishes three fundamental diagnostic dimensions: Self-Diagnosis, in which an individual artificial intelligence model or participating entity identifies and assesses problems within its own outputs, processes, decisions, or internal operational results; Mutual Diagnosis, in which two or more independent participants examine, evaluate, and identify potential problems in one another's outputs or activities; and Human Diagnosis, in which human individuals or groups participate in identifying, examining, interpreting, and assessing potential problems involving artificial intelligence systems or other participants. The framework does not assume that any single diagnostic dimension is universally sufficient or optimal. Instead, it introduces Universal Adaptive Diagnostic Selection and Switching, enabling an individual system or coordinating architecture to determine whether a diagnostic task should be conducted through self-based, mutual-based, human-based, or hybrid configurations. The diagnostic configuration can be dynamically modified according to task complexity, observed diagnostic performance, uncertainty, error severity, participant capabilities, disagreement, available evidence, resource constraints, and governance requirements. Furthermore, the framework distinguishes diagnosis from correction, diagnostic capability from diagnostic reliability, adaptation from switching, and simultaneous diagnostic combination from structural reconfiguration. It introduces a formal representation of diagnostic processes, a three-dimensional diagnostic configuration space, adaptive decision policies, dynamic transition mechanisms, diagnostic lifecycle, evaluation principles, and governance requirements. The framework supports a broad range of diagnostic functions, including factual error detection, logical inconsistency identification, reasoning assessment, hallucination detection, instruction-compliance diagnosis, security assessment, bias identification, procedural verification, decision evaluation, and system-performance diagnosis. To the best of our knowledge, this paper is the first to formally propose a unified, mechanism-independent theoretical framework that integrates Self-Diagnosis, Mutual Diagnosis, and Human Diagnosis with universal adaptive diagnostic selection, dynamic switching, combination, and reconfiguration within the broader USMHM architecture. Rather than asserting that the individual diagnostic mechanisms are unprecedented, this priority claim concerns their proposed unified organization and adaptive operational architecture. The framework provides a general conceptual foundation for designing and evaluating dynamically adaptive diagnostic systems involving heterogeneous artificial intelligence models and human participants.

Zenodo (CERN European Organization for Nuclear Research)
Good health and well-being
Explainable Artificial Intelligence (XAI)
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon