Artificial intelligence has for the first time predicted the reproductive behaviour of Yellowtail Kingfish by tracking their movements as part of new research revealed on #WorldOceanDay.
The new study published in Movement Ecology used machine learning algorithms to identify and distinguish between behaviours including courtship, feeding, escape, chafing, and swimming to showcase how technology can offer greater understanding of marine life.
Researchers tagged captive Kingfish and filmed their behaviour in tanks to identify the acceleration signatures and applied artificial intelligence to identify behaviour in free-ranging fish.
Flinders University PhD student, Thomas Clarke, in the College of Science & Engineering, says it’s the first study to use machine learning to identify spawning behaviours in wild Kingfish and demonstrates how artificial intelligence can be used to better understand reproductive patterns.
“Through direct observations of courtship and spawning behaviours, our findings provide the first study to predict natural reproduction of Yellowtail Kingfish, via the use of accelerometers and machine learning. This study builds on past work, which has been limited to direct observations, by applying such models on free-ranging data in a natural environment.”
“These findings offer the potential to identify naturally-occurring behaviours, which in the past have been only inferred through direct observations, or destructive approaches.”
Associate Professor Charlie Huveneers, who leads the Southern Shark Ecology Group at Flinders University, says the results showcase how cameras, tagging sensors and machine learning offer an unprecedented opportunity for animal conservation.
“Understanding reproductive behaviours is essential to predict population responses to environmental and fishing pressures, and to develop suitable and adaptable management strategies, where required.”