Abstract
<jats:p><p>A new approach to automating the practical application of intelligent assistants for analyzing, interpreting, and constructing texts and prompts in a given subject area based on their semantic content is proposed. This approach involves identifying semantic components and postulates in a given subject area, followed by calculating the associated semantic spectra based on the resulting semantic components. These spectra form the basis for quantitative comparisons based on specified criteria. The set of texts under study is formally considered as a space of discrete probability distributions, closed under the action of stochastic matrices, which represent the transformations of semantic profiles specified by prompts, meaningfully interpreted through decomposition in invariant subspaces of semantic components. The application of the proposed approach is illustrated within the framework of the cultural-historical psychology paradigm.</p></jats:p>