Train artificial intelligence to study the “song” of coral reefs

An artificial intelligence has been programmed by researchers at the University of Exeter to monitor the health of coral reefs by studying their sounds.

  • New research shows that artificial intelligence (AI) can track the health of coral reefs by analyzing the sounds and noises emitted by their constituent parts.
  • The University of Exeter has used the technology to monitor the progress of coral reef restoration.
  • The series of recordings successfully determined the health status of the corals 92 percent of the time.

What sound do corals make?? The University of Exeter in the UK decided to find out for scientific purposes. A team of researchers has created an AI program that can Determine coral health by listening to sounds that they make. These organisms, formed from colonies of polyps found at the bottom of the seas and oceans, produce complex Sounds and noises made by passing fish and other animals. Analyzing these special songs through AI allows researchers to obtain data useful in gauging coral health and launching restoration projects when necessary.

Corals and Science
A researcher places a water gun among coral reefs © Ben Williams – Exeter

Favorite choral songs

In the Study led by Professor Ben WilliamsThe algorithm was trained using an extensive database of Voices from both healthy and declining corals, allowing the machine to tell the difference. AI was then used to analyze the new recordings and was able to successfully determine the health of the corals 92 percent of the time. “Coral reefs face multiple threats Including Climate changeSo monitoring their health and the success of conservation projects is vital,” said Williams.

To date, one of the biggest difficulties has come from the fact that visual and acoustic analyzes of corals have been based on labor-intensive methods involving only specially trained scientists. Many coral reef creatures camouflage themselves, or are active at nightwhile the complexity of coral sounds made it difficult to determine coral health using individual recordings,” Williams explained. The Exeter team chose a technological approach inspired by machine learning to create software that can recognize the song of healthy corals. Our findings show that a computer can pick up patterns that are undetectable in the human ear. can you tell us Faster and more accurate, how do corals work. “

Coral reefs, degradation
Coral reefs that have suffered severe degradation © The Ocean Agency

The role of coral reefs and how to avoid their degradation

The recordings used in the study were made in the Indonesia, as some local coral reefs were severely damaged. The study authors state that the AI ​​method offers great opportunities for improving the monitoring of these important organisms. “This is a really exciting development. Voice recorders and AI can be used all over the world “To monitor the health of coral reefs, and to discover if attempts to protect and restore them are successful,” said Dr. Tim Lamont, co-author of the study. “In many cases, it’s easier and cheaper Spread hydroponic underwater on the reef and leave it there rather than inviting expert divers who frequently visit the reef to survey it – especially in remote locations.”

almost 25-50 percent of the world’s coral reefs have been destroyedand another 60 percent are threatened, According to the United Nations Environment Program. These ecosystems are vital sources of food and protect the coasts of island nations. 850 million people live within 100 km of coral reefs And reap the economic benefits from the nearest coral reefs. In the future, Williams is sure the use of AI can be extended to other locations around the world to help with restoration projects. “We want now Send recorders to all over the world: the Maldives, to the Great Barrier Reef, to Mexico, to a lot of different locations where we have partners that can collect similar data.”

Translated by

Licenza Creative Commons
Quest’opera and distribution with Licenza Creative Commons Attribuzione – Noncommercial – Non-derivative na Available 4.0 Internazionale.

Leave a Comment