Resumen
Representation and reasoning with temporal data is a well-researched problem in logic and computer science. Although many practical applications need the representation of inexact dates and reasoning with such representations, there is no standard developed methodology for it. In this paper, we propose a standard representation of inexact dates based on discrete probability distributions. Inspired by recent breakthroughs in natural language processing and information retrieval in embedding words as dense vectors we have developed a similar approach for representation and comparison of inexact dates.