Contributo in volume, 2019, ENG, 10.1007/978-3-030-25719-4

An Automatic System for Learning and Dialogue Based on Assertions

Umberto Maniscalco, Antonio Messina, Pietro Storniolo

ICAR Istituto di Calcolo e Reti ad Alte Prestazioni - CNR

When someone says a statement about a particular subject, we memorize the assertion and, implicitly, we can construct all the possible questions that have as a right answer to the assertion just heard. This means that, in this specific case, our learning process based on assertions subsists. When we read a book, we do nothing but learn through a succession of assertions. In this article, we present a system for automatically constructing a conversational agent, which uses only assertions to build the dialog engine. The whole architecture is based on the "Robot Operating System" (ROS), and the experiments were conducted using a humanoid robot.

Keywords

Learning, Robotics, Knowledge base, Conversational agent

CNR authors

Storniolo Pietro, Maniscalco Umberto, Messina Antonio

CNR institutes

ICAR – Istituto di calcolo e reti ad alte prestazioni