Model-Based Testability of Dynamical Metacognitive Systems: A Radio-Frequency Case
Proceedings of the AAAI Symposium Series · 06.11.2026 · 10 (1) · 97-104
DOI: 10.1609/aaaiss.v10i1.43395 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Symposium Series, online erschienen am 06.11.2026. Verfasst von Tyler Cody und 1 weiteren Autor:innen.
Abstract (Originalsprache)
Artificial metacognition introduces a moving assurance target: a system may alter its configuration, learned knowledge, or learning process while operating. Test obligations are obscured when metacognition is treated as synonymous with adaptation, transfer learning, or meta-learning. This paper develops a solution-method-agnostic basis for model-based testability of dynamical metacognitive systems. Extending an abstract systems theory of learning, the model separates object-level learning from self-observation, metacognitive state, control, and system change. A constructive separation establishes distinct roles for metacognition and meta-learning while identifying a nonempty intersection: metacognition is higher-order in reference, while meta-learning is higher-order in learning. Action-specific change contracts make the scope, authority, and provenance of adaptation explicit. Each test obligation is defined by a realizable intervention, retained-evidence projection, and verdict oracle, deriving tests of monitoring, diagnosis, control authority, change conformance, closed-loop recovery, safe abstention, and stability. A worked radio-frequency machine learning case instantiates the framework for automatic modulation classification under receiver, channel, interference, and offset changes. Together, these elements turn claims about evolving metacognitive behavior into bounded, auditable assurance obligations.
Zitieren
APA
Cody, T., & Beling, P. A. (2026). Model-Based Testability of Dynamical Metacognitive Systems: A Radio-Frequency Case. Proceedings of the AAAI Symposium Series, 10(1), 97-104. https://doi.org/10.1609/aaaiss.v10i1.43395
Vancouver
Cody T, Beling PA. Model-Based Testability of Dynamical Metacognitive Systems: A Radio-Frequency Case. Proceedings of the AAAI Symposium Series. 2026;10(1):97-104. doi:10.1609/aaaiss.v10i1.43395
BibTeX
@article{Cody2026,
author = {Cody, Tyler and Beling, Peter A.},
title = {Model-Based Testability of Dynamical Metacognitive Systems: A Radio-Frequency Case},
journal = {Proceedings of the AAAI Symposium Series},
year = {2026},
volume = {10},
number = {1},
pages = {97--104},
doi = {10.1609/aaaiss.v10i1.43395}
}
EUVIMED