Robot Concept

Epigenetic robotic is a new field that focuses on creating robots and modeling cognitive development that shows development of autonomous mental. For instance, robots have been implemented that generate visual discrimination behavior using large-scale neural networks which model early infant caregiver interaction using behavioral rules, and that explore the knowledge needed by infants to succeed in perceptual experiment of object permanence.

Given other diverse contributions to this new filed it seems an opportune time to synthesize a few core concepts from this corpus of research. It is one such core concept, ongoing emergence, which refers to the continuous integration and development of new skills. An agent was exhibiting ongoing emergence, in an autonomous manner motivationally, it will continue refine and develop its skills across development. The open-ended vision development is evident in recent work. In efforts to allow incrementally a mobile robot to progress through levels of behavior of increasingly sophisticated, in efforts to build robots that exhibit new behavior that becomes a precursor for successive development stages, and in efforts to reach robots exhibiting a successive emergence of behaviors in the progress of development of increasing processing complexity and power. Unfortunately, while human show clearly such long term progressions, epigenetic robots as they are typically designed to learn specific tasks or achieve particular behaviors.

A robot can learn using a developmental algorithm, via a self exploration process its behaviors repertoire and mental capabilities instead of being reprogrammed with the human concepts and the human body capabilities. Robots are proposed to discover even the most primitive behaviors through an exploration process.

A possible benefit of providing a developmental algorithm is avoiding specification of task objective for the robot. Instead to develop robotic to explore the tasks range that can be learned by a robot, given the control architecture and a specific developmental algorithm. It considers three mechanisms to be essential to developmental algorithms: anticipation, abstraction, and self motivation.