Verbal Description Focusing on Physical Properties of Real-world Environments

Abstract

In recent years, many prediction models for the real world have been proposed. We believe that when humans predict the real world, they capture the change points of observed objects and understand the recognized results as linguistic expressions. Based on this, in this study, we constructed a change-point prediction model based on changes in the observed object and verified its accuracy. In addition, the inferences made by the models were generated as language, and the predictions were verified.