Метод побудови нейросимвольного представлення ментальної моделі рішення інтелектуальної системи
Year: 2026
Type: article
Abstract: The paper addresses the problem of constructing mental models of intelligent system decisions for external users, which requires transparent explanations adapted to the application domain and the user’s expertise level. A generalised three-layer symbolic mental model is proposed, comprising a conceptual layer of decision abstractions, an explanation layer based on causal rules, and a fact base, which are linked by a vertical composition function that enables controlled explanation granularity. An algorithm for automated construction of the generalised model is developed using a domain ontology, causal reasoning paths, and user data, providing personalised explanations tailored to users’ expertise levels. The model supports explanation modes for novices, intermediate users, and experts, aligning the depth of detail with cognitive capabilities and verification requirements in high‑risk domains where trust in intelligent systems is essential.
Author: Лещинська, І.О.
Institution: Kharkiv National University of Radio Electronics
Language: Ukrainian
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7163852803
