Formal Definitions of Unbounded Evolution and Innovation Reveal Universal Mechanisms for Open-Ended Evolution in Dynamical Systems by Adams, A., Zenil, H., Davies, P. C. W., & Walker, S. I. (2017)

source
tags
Open-ended Evolution, Cellular automata

Summary

This paper defines two properties for dynamical systems which are claimed to be related to open-ended evolution: Unbounded evolution (UE) and Innovation (INN). The combination of these two properties makes a system open-ended according to this paper’s definition For such properties to be possible, a system has to be decomposed into two entities that interact with each other:

• an organism,
• an environment.

For the purpose of measuring “open-endedness”, the environment is ignored and the author wish to see if the organism can do more than what it could do if it was an isolated system.

Unbounded evolution

A system decomposed into an organism $$o$$ and an environment $$e$$ is said to exhibit UE if there exist a (state or rule) trajectory longer than the Poincaré recurrence time of the equivalent isolated organism.

Innovation

To exhibit innovation, a system $$o + e$$ must have a trajectory not contained within the possible trajectories of an isolated version of $$o$$.

Experiments with CA extensions

The authors experiment with a few extensions of the cellular automaton model to incorporate these organism/environment dynamics.