The Surprising Creativity of Digital Evolution: A Collection of Anecdotes from the Evolutionary Computation and Artificial Life Research Communities

By Joel Lehman, Jeff Clune, Dusan Misevic, C. Adami, Lee Altenberg, Julie Beaulieu, P. Bentley, Samuel Bernard, G. Beslon, David M. Bryson and 43 more. 2018

317 citations recorded by the source corpus.

Evolution provides a creative fount of complex and subtle adaptations that often surprise the scientists who discover them. However, the creativity of evolution is not limited to the natural world: Artificial organisms evolving in computational environments have also elicited surprise and wonder from the researchers studying them. The process of evolution is an algorithmic process that transcends the substrate in which it occurs. Indeed, many researchers in the field of digital evolution can provide examples of how their evolving algorithms and organisms have creatively subverted their expectations or intentions, exposed unrecognized bugs in their code, produced unexpectedly adaptations, or engaged in behaviors and outcomes, uncannily convergent with ones found in nature. Such stories routinely reveal surprise and creativity by evolution in these digital worlds, but they rarely fit into the standard scientific narrative. Instead they are often treated as mere obstacles to be overcome, rather than results that warrant study in their own right. Bugs are fixed, experiments are refocused, and one-off surprises are collapsed into a single data point. The stories themselves are traded among researchers through oral tradition, but that mode of information transmission is inefficient and prone to error and outright loss. Moreover, the fact that these stories tend to be shared only among practitioners means that many natural scientists do not realize how interesting and lifelike digital organisms are and how natural their evolution can be. To our knowledge, no collection of such anecdotes has been published before. This article is the crowd-sourced product of researchers in the fields of artificial life and evolutionary computation who have provided first-hand accounts of such cases. It thus serves as a written, fact-checked collection of scientifically important and even entertaining stories. In doing so we also present here substantial evidence that the existence and importance of evolutionary surprises extends beyond the natural world, and may indeed be a universal property of all complex evolving systems.

Subfields: Artificial Intelligence, Genetics, Sociology and Political Science, Reinforcement Learning

How researchers see this paper

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

In Artificial Intelligence

In Genetics

In Sociology and Political Science

In Reinforcement Learning

Citations in the corpus

Scalable agent alignment via reward modeling: a research direction (2018, background): “If the reward model only captures most aspects of the objective but not all of it, this can lead the agent to ﬁnd undesirable degenerate solutions (Amodei et al., 2016; Lehman et al., 2018; Ibarz et al., 2018).”
Supervising strong learners by amplifying weak experts (2018, background): “But aggressively optimizing that proxy can lead to pathological behavior (Lehman et al., 2018; Amodei and Clark, 2016; Amodei et al., 2016), an example of Goodhart’s Law.”
Scaling Laws for Reward Model Overoptimization (2022, background): “Previous work has shown numerous examples of such behavior in a wide variety of settings [Krakovna et al., 2020, Lehman et al., 2020].”
Unsolved Problems in ML Safety (2021, background): “For instance, Ofria’s [118] digital organisms evolved to detect when they were being monitored and would “play dead” to bypass the monitor, only to behave differently once monitoring completed.”
Emergent Tool Use From Multi-Agent Autocurricula (2019, background): “[66] Joel Lehman, Jeff Clune, Dusan Misevic, Christoph Adami, Lee Altenberg, Julie Beaulieu, Peter J Bentley, Samuel Bernard, Guillaume Beslon, David M Bryson, et al.”
Go-Explore: a New Approach for Hard-Exploration Problems (2019, background): “Finding bugs in games and simulators, as Go-Explore did, is an interesting reminder of the power and creativity of optimization algorithms [6], and is commercially valuable as a debugging tool to identify and fix such bugs before shipping simulators and video games.”
Learning when to Communicate at Scale in Multiagent Cooperative and Competitive Tasks (2018, background): “…match two teams which can communicate, can choose to not communicate at all (to prevent sharing strategies) or use dishonest signaling (to misdirect opponents) (Lehman et al., 2018) in order to optimize their own reward and handicap opponents; making it important to learn when to communicate.”
Reward Tampering Problems and Solutions in Reinforcement Learning: A Causal Influence Diagram Perspective (2019, background): “Indeed, Lehman et al. (2018) describes a large number of examples of reward functions gone wrong, often in humorous ways.”
The Surprising Creativity of Digital Evolution (2018, methodology): “We then curated over 90 submissions to create a “greatest hits” collection, currently available as a pre-print (Lehman, 2018).”
The Alignment Problem for Bayesian History-Based Reinforcement Learners∗ (2019, background): “Many more real-world examples of misspecified reward functions can be found in (Gwern, 2011; Irpan, 2018; Lehman et al., 2018).”
Towards Safe Artificial General Intelligence (2018, background): “Many more real-world examples of misspecified reward functions can be found in (Gwern, 2011; Irpan, 2018; Lehman et al., 2018).”
Understanding Agent Incentives using Causal Influence Diagrams. Part I: Single Action Settings (2019, background): “Another potential starting point is the wide range of surprising agent behaviors recorded by Lehman et al. (2018).”
AI-GAs: AI-generating algorithms, an alternate paradigm for producing general artificial intelligence (2019, background): “In fact, it is routine for researchers to be surprised by the strategies AI comes up with to optimize the objectives it is given, and often the strategies and behaviors it creates are not at all what the researcher intended [97].”
AGI Safety Literature Review (2018, background): “Gwern (2011), Irpan (2018), and Lehman et al. (2018) have many more examples.”