The ocean is awash in noise, and a new artificial intelligence technology may be able to assist scientists sort through it all in order to monitor and research marine species.
DeepSqueak is the name of the instrument, not because it measures dolphin cries at the depths of the ocean, but because it is based on a deep learning technique that was initially used to classify mice's ultrasonic squeals.
Researchers are already using the technique to analyze large datasets of marine bioacoustics.
Given that most of the ocean is beyond our physical grasp, underwater sound might aid in our understanding of where marine animals travel, their density and abundance, and how they interact.
Whale songs have already been used to detect an undiscovered population of blue whales in the Indian Ocean and a previously unknown species of beaked whale.
However, listening to ocean recordings and separating animal sounds from hours of waves, wind, and boat motors is time-consuming and difficult.
DeepSqueak can help with that. The system, which was recently presented at the Acoustical Society of America's 182nd Meeting, is meant to categorize underwater acoustic waves quicker and more correctly than any other approach available to date.
DeepSqueak scans through ocean sound data and generates heat maps depending on where and at what frequency particular acoustic signals are heard.
After that, the signals are directed to a single animal.
"Although we used DeepSqueak to detect underwater sounds, this user-friendly, open source tool would be useful for a variety of terrestrial species," says Elizabeth Ferguson, CEO and creator of Ocean Science Analytics, who presented the findings.
"The capabilities of call detection extend to frequencies below the ultrasonic sounds it was originally intended for. Due to this and the capability of DeepSqueak to detect variable call types, development of neural networks is possible for many species of interest."
The collection of marine acoustic noise has never been easier, but as hours of ocean soundscapes build up in databases throughout the world, scientists must find out how to make the most of it.
DeepSqueak might be a viable substitute to the human ear, allowing researchers to quickly identify and analyse noises all around the world.
During tests, the completely automated technology consistently detected the cries of certain marine mammals including humpback whales, delphinids, and fin whales.
It can also distinguish the sounds of these creatures from background noise, which is essential given that manmade sound is increasing the volume in the ocean.
DeepSqueak was initially released in 2019 as a tool for analyzing rats' and mice's extensive repertory of ultrasonic vocalizations.
The program was able to recognize a wide range of syllabic sounds after sifting through a series of squeaky recordings, and these brief mouse cries appear to be ordered in different ways depending on the context in which they are employed.
The findings might aid scientists in their research into how certain syllables and grammar in the mouse world express unique information. The noises a mouse makes under certain settings, for example, might be utilized to indicate fear, anxiety, or despair.
DeepSqueak might help scientists better understand the intricacies between animal vocalizations and behavior, especially in distant underwater underworlds where some of the planet's most elusive species swim.
The findings will be presented at the Acoustical Society of America's 182nd Annual Meeting.