The article presents an overview comparative analysis of the means and approaches to building relationships between the components of educational programs. The study used the methods of system analysis and analytical review, comparative analysis, as well as generalization of the identified limitations. The existing approaches to building relationships between the components of educational programs, such as ontological, graph, semantic, and based on natural language processing methods, are analyzed. Ontological models formalize the structure of an educational program, graph models identify interdisciplinary connections, and semantic and based on natural language processing models allow for the automation of matching student learning outcomes, discipline topics, and educational materials. However, these approaches have several limitations, such as dependence on manual markup, limited consideration of the actual content of discipline curricula, and limited analysis within a single course. The study concludes that a comprehensive approach is necessary and provides a foundation for further research on building relationships between the components of educational programs.
T. Saratova, Evgenia Vorontsova· Scientific and analytical jo...· 0 citations
The aim of this work is to analyze the architectural components of LLM-based multi-agent systems. The research method involves a comparative analysis of MetaGPT, Generative Agents, AutoGen, OrgAgent, MIRIX, and ChatDev architectures against a uniform set of criteria (role organization, memory, coordination, structured outputs, communication, quality control), followed by integration of the components into a single system and experimental evaluation on two task types in single-agent and multi-agent modes. It is found that on the examined tasks the multi-agent mode underperforms the single-agent baseline in quality (scores of 3,0 and 4,5 versus 9,0 on a ten-point scale). The architectural cause is identified: a mismatch between task type and the preconfigured action chain. A pattern is discovered: the integration of three verification loops – a failure detector, a compliance check, and a decision-making loop – produces a self-diagnosis capability not observed in any of the examined systems individually. The three loops jointly identified result unreliability and refused to deliver it. This property differs from known multi-agent systems diagnostic approaches where analysis is performed after task completion. The results may be applied when designing multi-agent systems for tasks where the cost of error is high.
T. Saratova, Aleksandr Apuhtin· Scientific and analytical jo...· 0 citations
