Social Robotics
speaker: Dr. Gerhard Sagerer, Rector, Bielefeld University, Germany event: Xu Guangqi Lecture Series date: 3:00 pm May 6, 2011 place: Grand Lecture Hall, SIBS Main Building
Measuring cognitive performance
Robots cannot be programmed for all possible situations, they need to be able to learn!
Statistical learning procedures need many training examples.
Learning from interaction: master and apprentice, learning as a social activity (a kind of communication)
Robocup tournament in Singapore: to teach a robot to pickup an item from a shop and bring it to the checkout
Interacting with robots
- subjects must not be computer scientists or familiar with robots
- feedback is very important
- emotions is crucial in human interactions
- feedback by emotion mimicry
Acoustic packaging: speech + motion segmentation
Flobi robot with facial expression
Social robotics
- getting humans to “believe” that the robots are more than they seem
- learning from interaction with humans