The evolved radio and its implications for modelling the evolution of novel sensors

By Jon Bird, Paul J. Layzell. 2002

65 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Ecology, Evolution, Behavior and Systematics, Evolutionary Computation

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 Ecology, Evolution, Behavior and Systematics

In Evolutionary Computation

Citations in the corpus

Concrete Problems in AI Safety (2016, background): “We may be able to ameliorate such problems by finding more efficient ways to exploit our limited oversight budget – for example by combining limited calls to the true objective function with frequent calls to an imperfect proxy that we are given or can learn.”
Computer Models of Creativity (2009, unknown): “In brief, a team at the University of Sussex were using a GA to evolve oscillator circuits—in hardware, not in simulation (Bird and Layzell 2002).”
Corrigibility (2015, methodology): “3 Simple examples of unforeseen solutions are familiar from contemporary AI systems: e.g., Bird and Layzell (2002) used ge-2 netic algorithms to evolve a design for an oscillator, and found that one of the solutions involved repurposing the printed circuit board tracks on the system’s motherboard…”
Formalizing Two Problems of Realistic World-Models (2015, background): “A common idiom named after Gregory Potemkin, who set up fake villages to impress Empress Catherine II. and Layzell (2002), which, tasked with designing an oscillating circuit, re-purposed the circuit tracks on its motherboard to use as a radio which amplified oscillating signals from nearby…”
Self-Modification of Policy and Utility Function in Rational Agents (2016, background): “Since highly intelligent agents may find unexpected ways of optimising a function (see e.g. Bird and Layzell 2002), it is important to use value functions such that any policy that optimises the value function will also optimise the behaviour we want from the agent.”
Towards Safe Artificial General Intelligence (2018, background): “Springer. isbn: 978-3-319-63702-0 • Tom Everitt and Marcus Hutter (2015b). “Analytical Results on the BFS vs. DFS Algorithm Selection Problem. Part I: Tree Search”. In: 28th Australasian Joint Conference on Artificial Intelligence, pp. 157–165 • Tom Everitt and Marcus Hutter (2015c). “Analytical Res”
Toward Idealized Decision Theory (2015, background): “gence, it is important for them to use highly reliable decision procedures. Verifying that a system works well in test conditions is not suﬃcient for high conﬁdence. Consider the genetic algorithm of Bird and Layzell (2002), which, if run on a simulated representation of a circuit board, would have ”
Aligning Superintelligence with Human Interests: A Technical Research Agenda (2015, background): “2 Highly Reliable Agent Designs Bird and Layzell (2002) describe a genetic algorithm which, tasked with making an oscillator, re-purposed the printed circuit board tracks on the motherboard as a makeshift radio to amplify oscillating signals from nearby computers.”
The alignment problem from a deep learning perspective (2022, background): “…in a way that shifts the distribution of future episodes; or they could learn strategies that depend on data from unintended input channels (as in the case of an evolutionary algorithm which designed an oscillator to make use of radio signals from nearby computers [Bird and Layzell, 2002]).”
Defining and Characterizing Reward Hacking (2022, unknown): “…a racing game (Clark and Amodei, 2016); an evolved circuit listens in on radio signals from nearby computers’ oscillators instead of building its own (Bird and Layzell, 2002); universities reject the most qualified applicants in order to appear more selective and boost their ratings (Golden,…”
How the body shapes the way we think - a new view on intelligence (2006, methodology): “In another landmark experiment conducted by Jon Bird and Paul Layzell (Bird and Layzell, 2002), also of Sussex University, a circuit was evolved to produce an oscillatory signal without using any internal clock.”