Artificial Intelligence in Swarming and Aggregation
Artificial intelligence can aid researchers when testing the behaviours of an individual on how they take part in group movements and how they are coordinated; this often can be done through computer simulations. These simulations can detect the rules which play part in controlling the behaviours of small worms that feed on bacteria in large groups.
The computer tracks the movement of the worms through microscope images which can be viewed in time lapsing movies. The worms move in clusters over food and when it begins to disappear, the cluster moves together in search for more food. These new techniques have uncovered the behaviours of the individual worm inside the cluster and was able to track the movements which gives the researchers more insight as to how these worms behave. By understanding those behaviours, scientists will benefit in a much wider range of applications; as this can be applied to a single bacteria to bigger organisms like birds and animals. The results will aid the scientists understand whether there is a universal rule to predict group and individual behaviours.
This AI technique use fluorescent imaging and it maps out a phase diagram which entails behaviour phenotypes by showing the modulation of cluster-edge reversals as well as a dependency on density switching between speeds which are sufficient to give aggregation, while the past experiments had focused on compact clusters. Aggregation has often been categorized by the fraction of worms inside a given cluster, however newer AI techniques have added dynamic swarming phenotype of aggregation. Swarming is the collective movements of a cluster, which was often used in past experiments. This proccess takes a long timescale which was never obvious or detected through manual observations. Recently however, thanks to AI modern advances, the long timescale is no longer a problem as the computer can now replace it with a short time lapse video revealing more accurate results.