Systemic And Hole Semantics In Human-Machine Language Interfaces

Abstract

This paper concerns the progress of earlier related work based in an experiment to enhance the intelligence of robotic systems, with the aim of achieving more language communication capabilities between humans and robots. For that purpose, a simple artificial language was used and a systemic model of language communication. The robot, which was implemented through a computer simulation, is initially taught both the artificial language and the systemic model of human communication. The main goal of improvement was to allow a robot using the implied information of incoming sentences and ask questions whenever the given instructions are not complete. This process was based on the voids of an output data structure that corresponds to a systemic model of language communication, augmented with hole semantics, as grammar formalism, in order to handle partial/incomplete information. As expected, the improvement of learning abilities was observed. Some problems were also noticed in the process that had to be handled. This research comes to redefine the project in relation to these problems, as well as the final form of the language that was planned to be used. The usage of a dictionary of approximately 200 words in Greek natural language, the interchange of the programming language from Java to Python, as well as a new word learning procedures, are the new points in this paper. New words help the robot to expand the initial dictionary. The learning procedures start at the moment that a unknown word, included in a sentence addressed to the robot, arrives in its hearing input system. In addition, Python is an evolving programming language in various artificial intelligence's applications. Finally, the natural language helps in a more direct and friendly communication with most users, in free expressions with fewer restrictions.