How can an autonomous robot acquire the meaning of new words ?
The Talking AIBO
project investigated the mechanism that enables an autonomous robot to
learn how to use words in appropriate situations based on its social
and perceptual history. An enhanced version
of AIBO -- Sony's four-legged robot -- tried to construct from scratch
the meaning of simple words uttered by humans. These words concerned the
presence of objects (ball, red, etc), the behavior of the robot (walk,
sit) and the robot's body part (leg, head). These experiments showed the
importance of grounding each word in its social and perceptual context.
Selected Publications
Steels, L. and Kaplan, F. (2000) AIBO's first words : the social learning of language and meaning, Evolution of Communication 4 (1).
Kaplan, F., Oudeyer, P-Y., Kubinyi, E. and Miklosa, A. (2002) Robotic clicker training, Robotics and Autonomous Systems, 38(3-4):197-206. [pdf]
)
