Artificial intelligence in psychodiagnostics: cognitive states in a digital educational environment

<p>The article discusses the task of building multimodal AI models for diagnosing the cognitive state of students (concentration, fatigue, stress) in digital educational environments. The necessity of transition from traditional methods of psychodiagnostics to automated systems based on natural language processing, computer vision and behavioral analysis is substantiated. A mathematical model based on the CNN-LSTM hybrid architecture with the adaptation of parameters to individual cognitive profiles is proposed. The structure of the model is described, recommendations for its construction and integration into the digital educational infrastructure are given. The problems of interpretability, privacy, and sustainability of such models, as well as the prospects for their application, are discussed.</p>