nature & planetPhys.org

Machine learning reveals how often fish schooling evolved—and repeatedly disappeared

from Phys.org

Scientists have used new machine learning techniques to investigate a topic no one has ever found a way to study: the evolution of fish moving together in schools across thousands of species. The new paper, published Nov. 6 in the Proceedings of the National Academy of Sciences, is among the largest studies of the evolution of any behavior ever conducted in any group of species, according to the authors. The research shows that behaviors, which have previously been difficult to quantify, can be studied at a large evolutionary scale.

explained

Machine learning is a field of artificial intelligence focused on statistical algorithms that learn from data and generalise to unseen data, performing tasks without being explicitly programmed. It rests on statistics and mathematical optimisation, and deep learning's neural networks have surpassed many earlier approaches. Because of this dominance, the terms machine learning and artificial intelligence are now often used interchangeably, though they are not synonymous.

Background adapted from Wikipedia: Machine learning ↗ · text CC BY-SA 4.0
how it works

Schooling is costly as well as useful: it protects from predators and saves energy, but it also means competing for food and spreading disease. Whether the benefits win depends on a species' habitat and lifestyle, so it can be gained or lost.

so what

Behaviour leaves no fossils, so it is hard to study in evolutionary time. If machine learning can score behaviour at this scale, researchers can ask the same kind of question about migration, mating displays and other traits that are hard to measure.

more on this

Illustrative photo · lifefish, CC BY 3.0, via Wikimedia Commons