DocumentCode :
565795
Title :
How people talk when teaching a robot
Author :
Kim, Elizabeth S. ; Leyzberg, Dan ; Tsui, Katherine M. ; Scassellati, Brian
Author_Institution :
Dept. of Comput. Sci., Yale Univ., New Haven, CT, USA
fYear :
2009
fDate :
11-13 March 2009
Firstpage :
23
Lastpage :
30
Abstract :
We examine affective vocalizations provided by human teachers to robotic learners. In unscripted one-on-one interactions, participants provided vocal input to a robotic dinosaur as the robot selected toy buildings to knock down. We find that (1) people vary their vocal input depending on the learner´s performance history, (2) people do not wait until a robotic learner completes an action before they provide input and (3) people naïvely and spontaneously use intensely affective vocalizations. Our findings suggest modifications may be needed to traditional machine learning models to better fit observed human tendencies. Our observations of human behavior contradict the popular assumptions made by machine learning algorithms (in particular, reinforcement learning) that the reward function is stationary and path-independent for social learning interactions. We also propose an interaction taxonomy that describes three phases of a human-teacher´s vocalizations: direction, spoken before an action is taken; guidance, spoken as the learner communicates an intended action; and feedback, spoken in response to a completed action.
Keywords :
feedback; human-robot interaction; learning (artificial intelligence); speech processing; affective vocalization; completed action; feedback; human behavior; human teachers; human-teacher vocalization; intended action; interaction taxonomy; learner communication; learner performance history; machine learning algorithm; machine learning model; reinforcement learning; reward function; robot teaching; robotic dinosaur; robotic learners; social learning interaction; unscripted one-on-one interaction; vocal input; Buildings; Dinosaurs; Humans; Learning; Machine learning; Robots; Training; Human-robot interaction; affective input; affective vocalization; naïve teaching; reinforcement learning; social learning;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Human-Robot Interaction (HRI), 2009 4th ACM/IEEE International Conference on
Conference_Location :
La Jolla, CA
ISSN :
2167-2121
Print_ISBN :
978-1-60558-404-1
Type :
conf
Filename :
6256043
Link To Document :
بازگشت