Refinement of the mental model of a solution based on input data augmentation in the task of generating explanations in an intelligent system

The subject of the research is the process of constructing a mental model of a solution for an intelligent system user. The aim is to develop an approach to refine the user's mental model based on the principle of input data augmentation to account for both positive and negative aspects of the solution. Objectives: structuring the principle of input data augmentation; development of a method for refining the mental model of an intelligent system user based on input data augmentation. The scientific novelty of the obtained results is as follows. The principles of constructing mental models have been improved, taking into account the peculiarities of generating explanations in intelligent systems. A method for refining the mental model based on the analysis of user feedback is proposed, which allows supplementing the model with negative characteristics of the solution. The method includes stages of analyzing user feedback, assessing the importance of identified solution characteristics, supplementing the mental model with negative characteristics, and balancing positive and negative aspects of the solution. The use of this method in constructing explanations enables the formation of more comprehensive and balanced explanations that consider both the advantages and limitations of the intelligent system's solution. Experimental verification of the proposed method demonstrated the possibility of automated augmentation of the mental model of the solution based on the analysis of intelligent system user feedback.