Distributed Representations, Simple Recurrent Networks, And Grammatical Structure
Machine Learning - Connectionist approaches to language learning
An introduction to genetic algorithms
An introduction to genetic algorithms
Artificial Life: An Overview
Evolutionary Robotics: The Biology, Intelligence, and Technology of Self-Organizing Machines
Evolutionary Robotics: The Biology, Intelligence, and Technology of Self-Organizing Machines
Hi-index | 0.00 |
This paper shows how spatial cognition and the ability to orient in a closed arena can also emerge in artificial organisms with no sensory apparatus. The control systems (Artificial Neural Networks) of some populations of simulated robots have been evolved in an experimental set-up that is usually used to study spatial cognition of real organisms. Some robots were endowed with a normal perceptive apparatus (infrared sensors and on-board camera), some others had no means of getting information from environment. The control systems of this latter class of robots received stimulation from self-generated input: the feedback of their motor action or the activation of an internal clock. Both kinds of robot learnt some strategies similar to the ones observed in natural organisms. The robots without sensory apparatus displayed a greater amount of micro-behaviour (number of activation patterns of motor output) compared with robots with a normal perceptual system. The amplification of behavioural repertory allowed them to understand the environmental structure even if they couldn't perceive it.