Research on the Dynamic Mechanism of the Four-Dimensional Model of Self-Evaluation and Learning Elements under the Empowerment of Generative AI

Abstract

Self-evaluation in self-regulated learning directly impacts learning outcomes, and generative artificial intelligence (AI) offers new opportunities for optimizing this ability. This study constructs a self-evaluation model encompassing four dimensions: formative, summative, metacognitive, and affective. The dynamic interaction between this model and learning elements (motivation, strategy, behavior, and cognition) is analyzed in depth. Through coupling relationship analysis, exploration of influence pathways, and interpretation of feedback mechanisms, the bidirectional interaction between the four-dimensional model and learning elements is revealed. Based on this, an adaptive optimization strategy for this dynamic interaction mechanism is proposed, providing theoretical support and practical approaches for building an intelligent learning ecosystem. The research adopts a combination of literature review and theoretical modeling to construct an analytical framework, and verifies the practical application value of the model based on empirical data from typical cases such as iFlytek Smart Classroom (covering 38000 schools, generating 120 million learning data per day), Squirrel AI Intelligent Adaptation System (serving over 4 million students, with a granularity of 12000+knowledge points), and Xueersi Online School Intelligent Teaching Platform (with a peak concurrent user volume of 500000+per day).