Identifying Necessary Conditions for Open-Ended Evolution through the Artificial Life World of Chromaria

By Lisa B. Soros, Kenneth O. Stanley. 2014

99 citations recorded by the source corpus.

A full understanding of open-ended evolutionary dynamics remains elusive. While artiﬁcial life worlds have been proposed to study such dynamics and tests have been devised to try to detect them, no theory yet has enumerated the key conditions that are essential to inducing them. The aim of this paper is to further such an understanding by hypothesizing four conditions that are essential for open-ended evolution to prosper. Of course, any such conditions must be satisﬁed by nature (the clearest example of an open-ended domain), but we do not know the scope or range of possible worlds that could achieve similarly impressive results. To complement the hypothesized conditions, a new artiﬁcial life world called Chromaria is introduced that is designed explicitly for testing them. Chromaria, which is intended to deviate from Earth in key respects that highlight the breadth of possible worlds that can satisfy the four conditions, is shown in this paper to stagnate when one of the four conditions is not met. This initial controlled experiment thereby sets the stage for a broad research program and conversation on investigating and controlling for the key conditions for open-ended evolution.

Subfields: Genetics, Mechanical Engineering, Sociology and Political Science

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Genetics

In Mechanical Engineering

In Sociology and Political Science

Citations in the corpus

Emergent Tool Use From Multi-Agent Autocurricula (2019, background): “More recent work attempted to formulate necessary preconditions for open-ended evolution [25, 26].”
Paired Open-Ended Trailblazer (POET): Endlessly Generating Increasingly Complex and Diverse Learning Environments and Their Solutions (2019, background): “…evolving individuals [30], MCC introduce a new kind of coevolution that evolves two interlocking populations whose members earn the right to reproduce by satisfying a minimal criterion [25, 26, 53] with respect to the other population, as both populations are gradually shifting simultaneously.”
AI-GAs: AI-generating algorithms, an alternate paradigm for producing general artificial intelligence (2019, background): “Much work has occurred in the fields of computational evolutionary biology, artificial life, and neuroevolution attempting to investigate how to create an open-ended complexity explosion like the one that occurred with living animals on Earth [99, 172, 13, 88, 138, 17, 162, 2, 183, 70, 86, 85, 167, ”
Enhanced POET: Open-Ended Reinforcement Learning through Unbounded Invention of Learning Challenges and their Solutions (2020, background): “…pursuing different kinds of solutions simultaneously) and optimization (e.g. giving one arguably “best” solution) (Brant & Stanley, 2017; Lehman & Stanley, 2011b; Mouret & Clune, 2015; Pugh et al., 2016; Soros & Stanley, 2014), as those solely focusing on optimization often lead to convergence.”
POET: open-ended coevolution of environments and their optimized solutions (2019, background): “Unlike conventional coevolutionary algorithms that are usually divided between competitive coevolution [14] and cooperative coevolution [62], MCC introduces a new kind of coevolution that evolves two interlocking populations whose members earn the right to reproduce by satisfying aminimal criterion ”
Open-Ended Evolution: Perspectives from the OEE Workshop in York (2016, unknown): “of research along these lines has recently been initiated by Soros and Stanley [44].”
Minimal criterion coevolution: a new approach to open-ended search (2017, methodology): “However, a key factor in Chromaria’s success is that organisms interact with each other to satisfy the MC, which allows its difficulty to vary organically over time as the population changes.”