Bengaluru: Researchers at the Indian Institute of Information Technology Bengaluru (IIIT-B) have developed a robotic model that they say understands humans and interacts with them based on emotions.
Models that exist in the market predict engaging behaviour directly from multi-modal features and do not incorporate personality inferences or any theories of interpersonal behaviour in human-human interactions, according to the researchers.
A research paper by Soham Joshi, Arpitha Malavalli and Shrisha Rao was published recently in PLOS, an open-access publisher.
“We want the robot to show cognitive improvement in a sensible way, not in a completely fixed manner. If you have a query as a person and attitude, the response from the human is different. What we want to see is an automated system which more accurately models the engagement and behaviour of human interaction with them and adjusts their responses accordingly,” Rao, the faculty supervisor for the project, said.
Rao further explained that this is a pipeline way in which long-existing psychological theories have been brought into the field of Artificial Intelligence (AI).
“We’ve used classical psychological theories about human personality engagement and so on and help the robot or system to build a better model and how to predict human engagement and adjust its behaviour,” he added.
