Recent site activity

Project Abstract

The Emergence of Artificial Culture in Robot Societies

A profound question that transcends disciplinary boundaries is "how can culture emerge and evolve as a novel property in groups of social animals?" We can narrow that question by focussing our attention on the very early stages of the emergence and evolution of simple cultural artefacts; the transition, as it were, from nothing recognisable as culture, to something (let us call this proto-culture). This project aims to address and illuminate that question in a radical and hitherto inconceivable new way by building an artificial society of embodied intelligent agents (real robots), creating an environment (artificial ecosystem) and appropriate primitive behaviours for those robots, then free running the artificial society. Even with small populations (a few tens) of relatively simple robots we will, in a short time, see a very large number of interactions between robots. The inherent heterogeneities of real robots, and the noise and uncertainty of the real world, vastly increase the space of possibilities and the scope for unexpected emergence in the interactions between robots.

In this project we will aim to create the conditions and primitives in which proto-culture can emerge in a robot society. Robots will, for example, be able to copy each other's behaviours and select which behaviours to copy. Behaviours (memes) will mutate because of the noise and uncertainty in the real robots' sensors and actuators, and successful memes will undergo multiple cycles of copying (heredity), selection and variation (mutation). Furthermore we will introduce a bi-phased approach in which we alternate between real-time (with real physical robots) in which the emergence, selection and refinement of these discrete behavioural artefacts takes place; with evolutionary time, in which we run a genetic algorithm (GA) process to grow and evolve the robots' controllers so that the behaviours and premiums associated with the emerging memes become hard-wired into the robots' (neural) controllers.

In this way we hope to see the emergence of interesting behavioural artefacts that, we hope, will be qualitatively and quantitatively distinct from those present at the beginning. Of course the behavioural artefacts that emerge and evolve, that we hope to identify as proto-cultural analogues, will not be human but decidedly robotic. We do not expect these artificial memes to have any meaning in a human cultural context; rather, they will be meaningful only within the closed context of this artificial society (an exo-culture). A significant challenge for this project will therefore be to identify and interpret these patterns of behaviour as evidence for an emerging exo-culture; the challenge is hermeneutic - what means will we be able to develop by which we can identify/recognise meaningful/cultural behaviour; and, then, what means might we go on to develop for interpreting/understanding this behaviour and/or its significance?