Google has developed a brand new AI tool to assist marine biologists higher understand coral reef ecosystems and their health, which may aid in conversation efforts. The tool, SurfPerch, created with Google Research and DeepMind, was trained on 1000’s of hours of audio reef recordings that allow scientists studying the reef to have the opportunity to “hear reef health from the within,” track reef activity at night, and track reefs which can be in deep or murky waters.
The project began by inviting the general public to take heed to reef sounds via the online. Over the past yr, visitors to Google’s Calling in our Corals website listened to over 400 hours of reef audio from sites around the globe and were told to click after they heard a fish sound. This resulted in a “bioacoustic” data set focused on reef health. By crowdsourcing this activity, Google was in a position to create a library of latest fish sounds that were used to fine-tune the AI tool, SurfPerch. Now, SurfPerch could be quickly trained to detect any recent reef sound.
“This permits us to investigate recent datasets with much more efficiency than previously possible, removing the necessity for training on expensive GPU processors and opening recent opportunities to know reef communities and conservation of those,” notes a Google blog post in regards to the project. The post was co-authored by Steve Simpson a professor of Marine Biology on the University of Bristol within the U.K., and Ben Williams, a marine biologist on the University College London, each who study coral ecosystems with focuses on areas like climate change and restoration.
What’s more, the researchers realized they were in a position to boost SurfPerch’s model performance by leveraging bird recordings. Although bird sounds and reef recordings are very different, there have been common patterns between bird songs and fish sounds that the model was in a position to learn from, they found.
After combining the Calling Our Corals data with SurfPerch in initial trials, researchers were in a position to uncover differences between protected and unprotected reefs within the Philippines, track restoration outcomes in Indonesia, and higher understand relationships with the fish community on the Great Barrier Reef.
The project continues today, as recent audio is added to the Calling in Our Corals website, which is able to help to further train the AI model, Google says.
