Social interactions of Hyla japonica helps to shape artificial intelligence.
Scientists are increasingly studying groups of animals, ranging from ant colonies to shoals of fish, flocks of birds and even groups of frogs, to gain insight into how the natural world is organized. This is a key feature of a field of investigation now known as ‘swarm intelligence’ that seeks to identify and then replicate such pathways as part of the development of artificial intelligence systems.
Tanaka Juuyoh
The ‘desynchronised’ calls of the Japanese tree frog has inspired the development of an artificial intelligence algorithm.
Avoiding Direct Competition
A major breakthrough in this field has just emerged from studies involving the Japanese tree frog (Hyla japonica). Researchers at the Polytechnic University of Catalonia in Spain observed how males of this species have evolved a system of not calling simultaneously with a rival, in the presence of a potential mate.
This provides the female with the opportunity to distinguish the individuals on this basis, which would otherwise not be possible if they were close-by and vocalizing at the same time. Males therefore recognize the need to “desynchronize” their calls, to use the jargon, so they do not call together.
"Since there is no system of central control organizing this "desynchronization," the mechanism may be considered as an example of natural self-organization," explains Christian Blum, researcher at the Polytechnic University of Catalonia in Spain.
Interpreting and Using This Method
Working with his colleague Hugo Hernández, the pair then used this type of calling behavior – carried out without blocking – to create a mathematical model called an algorithm. Its function is to assign colors to network nodes, which is a technique that can be used to develop more efficient wireless (WiFi) networks.
The algorithm assigns colors to network nodes, with the aim being to ensure that each pair of these network nodes when connected is not the same color, rather like the way that the frogs do not call together. The ultimate aim is to find a solution that uses the fewest colors.
The scientists hope to address a problem that is increasingly significant in the modern world. “We want to optimize modern wireless networks with no predetermined structure using techniques for reducing losses in information packages and energy efficiency improvement,” explains Blum.
Reference
Hugo Hernández, Christian Blum. Distributed graph coloring: an approach based on the calling behavior of Japanese tree frogs. Swarm Intelligence, 2012; 6 (2): 117 DOI: 10.1007/s11721-012-0067-2
