INRIA logo

Speaker: Serena Ivaldi, research scientist at Larsen team, INRIA Grand-Est, Nancy, France.

Abstract: In this talk I will review some of our recent experiments about human-robot interaction, in particular with the iCub and the Kinova Jaco robots interacting with people that are not experts in robotics. I will discuss why machine learning is important to make the robot learn objects and integrate human feedback; why we need to take into account multimodal signals, to cope with physical interaction, but also to enrich the social communication and make it intuitive for humans; then why we need to consider also a social dimension, as people interacting with robots often do not only see a machine but also an agent, with its own intelligence and decision capabilities, that poses challenges for acceptance and trust.