Searching for Emergent Representations in Evolved Dynamical Systems
Thomas Hope, Ivilin Stoianov and Marco Zorzi
Simulation of Adaptive Behavior 2006 (SAB 2006)
Rome, Italy, 25-29 September 2006
Summary
This paper reports an experiment in which artificial foraging agents with dynamic, recurrent neural network architectures, are "evolved" within a simulated ecosystem. The resultant agents can compare different food values to "go for more," and display similar comparison performance to that found in bio-logical subjects. We propose and apply a novel methodology for analysing these networks, seeking to recover their quantity representations within an Ap-proximationist framework. We focus on Localist representation, seeking to in-terpret single units as conveying representative information through their aver-age activities. One unit is identified that passes our "representation test", repre-senting quantity by inverse accumulation.