Google has developed a new AI tool to help marine biologists better understand coral reef ecosystems and their health, which could aid communication efforts. SurfperchThe project, created in collaboration with Google Research and DeepMind, was trained on thousands of hours of audio reef recordings, allowing scientists studying reefs to “listen to reef health from the inside,” track reef activity at night, and spot reefs in deep or murky waters.
The project began by inviting people to listen to the sounds of the reef via the web. Over the past year, visitors to Google have Calling our Coral website Over 400 hours of reef audio from sites around the world were listened to and they were told to click when they heard a fish sound. This resulted in a “bioacoustic” data set focused on reef health. By crowdsourcing this activity, Google was able to create a library of new fish sounds that was used to improve the AI tool, SurfPerch. Now, SurfPerch can be quickly trained to detect any new reef sound.
“This allows us to analyse new datasets with much greater efficiency than previously possible, eliminating the need for training on expensive GPU processors and opening up new opportunities for understanding reef communities and their conservation,” notes a google blog post About the project. This post is co-authored by Steve Simpson, a professor of marine biology at the University of Bristol, UK, and Ben Williams, a marine biologist at University College London, both of whom study coral ecosystems, focusing on areas such as climate change and restoration.
In addition, the researchers felt they were able to enhance the performance of the surfperch model Bird Recordings. They found that although the bird calls and the reef recordings differed greatly, there were some common patterns between the bird songs and the fish calls, which the model was able to learn.
Calling our corals in the beginning after combining data with SurfPerch TrialsThe researchers were also able to highlight differences between protected and unprotected reefs in the Philippines, track restoration results in Indonesia, and better understand relationships with the fish community on the Great Barrier Reef.
The project continues to this day, as new audio is added. Calling on our Coral websiteGoogle says this will help train AI models more efficiently.
google-looks-to-ai-to-help-save-the-coral-reefs-techcrunch
