Feel.Back.Loop

Le résumé fourni par la source

Future working environments will require humans to closely collaborate with artificial intelligence and robotic systems. In particular, robots are increasingly expected to incorporate emotional sensitivity while becoming more adaptive in their behavior. Future human-robot interaction will go beyond simple task structures, and robots may need to be configured or even collaboratively trained by their human partners, who are not necessarily experts in machine learning. The artwork Feel.Back.Loop addresses these questions through an interactive installation. It presents an abstract robotic system in the form of a tail that is continuously trained by visitors using reinforcement learning. Spectators’ emotional reactions are the basis of the robot’s reward function, and multiple encounters build up a policy for its movements over time, evolving from initially random actions into more consistent behavioral patterns. By embedding reinforcement learning within an interactive artwork, the project not only explores adaptive human–robot collaboration but also foster reflection and discussions about the role of human participation, training, and affect in shaping intelligent systems in future workplaces.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.