What is it about?
This work presents a reinforcement learning algorithm that enables a simulated robot to adapt quicker to rapid changes in human feedback during social interaction.
Featured Image
Why is it important?
Enabling robots to have social interactions with humans require them to be able to flexibly and rapidly adapt to changes in subtle social feedback, like human attention towards the robot or towards the task, which vary constantly and can be perturbed by external events.
Perspectives
The perspective of this work is to then test the reinforcement learning algorithm during real social interactions with humans, and in particular children, to evaluate its efficiency in the real world.
Mehdi Khamassi
Read the Original
This page is a summary of: Adaptive reinforcement learning with active state-specific exploration for engagement maximization during simulated child-robot interaction, Paladyn Journal of Behavioral Robotics, August 2018, De Gruyter,
DOI: 10.1515/pjbr-2018-0016.
You can read the full text:
Resources
Contributors
The following have contributed to this page
