Cognitive situated agents learn to name actions

Authors:
Julien Poudade;Lionel Landwerlin;Patrick Paroubek
Affiliations:
LIMSI-CNRS, Orsay, France, email: {poudade,landwerlin,pap}@limsi.fr;LIMSI-CNRS, Orsay, France, email: {poudade,landwerlin,pap}@limsi.fr;LIMSI-CNRS, Orsay, France, email: {poudade,landwerlin,pap}@limsi.fr
Venue:
Proceedings of the 2006 conference on ECAI 2006: 17th European Conference on Artificial Intelligence August 29 -- September 1, 2006, Riva del Garda, Italy
Year:
2006

Citing 5
Cited 1

Cyc: toward programs with common sense

Communications of the ACM
Intelligence without representation

Artificial Intelligence
Synthetic brain imaging: grasping, mirror neurons and imitation

Neural Networks - Special issue on the global brain: imaging and modelling
Is there something out there?: Inferring space from sensorimotor dependencies

Neural Computation
Spontaneous evolution of linguistic structure-an iterated learningmodel of the emergence of regularity and irregularity

IEEE Transactions on Evolutionary Computation

Emergence of language: hidden states and local environments

Proceedings of the 2008 conference on Information Modelling and Knowledge Bases XIX

Quantified Score

Hi-index	0.00

Visualization

Abstract

We present experiments in which a population of situated autonomous cognitive agents learns by means of “language games” a common lexicon of symbols that designates basic motions (translation, rotation) in a 2D world. We study the quality of the lexicon learnt when we give ours agents a cognitive functionality akin to working memory, and we explain how they are able to agree on a set of common names for basic actions, if their world representation is built using sensorimotor contingencies informations.