A probabilistic approach to learning a visually grounded language model through human-robot interaction | IEEE Conference Publication | IEEE Xplore