Meeting Abstract
Evolvability is the theoretical concept that attempts to capture the potential for populations to undergo future evolutionary change. Genetic variation, gene-to-phenotype mapping, and phylogenetic inertia are all thought to be important. Grounding evolvability in quantitative metrics and testable hypotheses is the goal of our research. It has been proposed that conditions of the agents and the environment that favor the evolution of modular neural connectomes will also create populations with enhanced evolvability. We test this hypothesis using Tadros, physically embodied robots inspired by the tadpole larvae of ascidians, which are autonomous surface swimmers. Tadros have a genome that encodes for a neural network that connects sensors to the tail-flapping motor, coordinating motor responses to changes in sensory input. We examine the structure and function of this evolving connectome as we apply selection for enhanced light gathering to a population of Tadros. Funded by the National Science Foundation (INSPIRE, Special Projects 1344227).
