Minimal and Necessary Conditions for the Emergence of Species-Specific Recognition Patterns

A simple mechanism is presented for the emergence of recognition patterns that are used by individuals to find each other and mate. The genetic component determines the brain of an individual, a machine learning architecture which is then used to transmit knowledge. Thanks to the interactions between the genetic and the knowledge parts the agents get to use species-specific recognition patterns, starting from an initial condition where the species are not distinguishable. Several machine learning architectures are investigated, as well as the influence of space and asynchronous genetic algorithm operations. Agents selecting each other for mating based on their limited recognition capacities is all that is needed for the emergence of species-specific recognition patterns: the transition between symbols to sequences with an intrinsic role within the species.

• Type: Conference
• Year: 2007
• Where: AE
• Authors: Nicolas Brodu
• Keywords: AE 2007, Artificial Intelligence, Machine Learning Architectures, Recognition Patterns, Species-specific Recognition Patterns