Have you ever wondered why your virtual assistant doesn't understand your questions? Or why did your GPS take you to the side road instead of the highway? In a study published April 21 in the journal iScience , Sicilian researchers designed "Pepper", a robot that "thinks aloud" so that users can listen to its thought process and better understand the motivations and decisions of the robot.
The research, conducted at the RoboticsLabs at the Engineering Department of the University of Palermo , in the context of computational models of language, information technology and robotics, demonstrates how through the inner speech the robot is able to improve the achievement of its goals .
The study is an important step towards man-machine communication . The idea that a robot can have some sort of thinking is amazing and this can help the interaction, creating the right level of trust. The new Robot model learns new things like children do, arousing empathy in humans.
At the table with the Pepper robot
The appearance of the robot used in the experiment is that of Pepper, a small humanoid with large eyes used in many laboratories around the world. The software that allows him to think aloud is unique and usable on other humanoids. The novelty is that, when preparing to execute a command, the robot gives voice to the thoughts that accompany its actions and does so by slightly lowering the tone of the voice.
For example, when asked to place the napkin on the table in a position that seems contrary to what the label says, the robot says in a low voice:
“ The position must be on the plate and not on the table. I think it's a bit confusing. I want to be sure. I'll ask him again "
Once confirmed, think aloud again:
“ Um, this situation bothers me, I never break the rules. I'm afraid to let him down, so I'll do what he wants "
Pepper and his inner language
Inner language is common in people and is used to gain clarity, seek moral guidance, and evaluate situations in order to make better decisions.
Scientists have found that, with the help of inner language, Pepper is better at solving dilemmas . Comparing Pepper's performance with and without inner language, they found that the robot had a higher task completion rate when engaging in dialogue with itself. Thanks to the inner language, Pepper has surpassed the functional and moral requirements of international standards for collaborative robots .
Curated by Antonino Pagano