Harvard Scientists Develop Innovative Bio-Logger to Decipher Sperm Whale Language

Say you want to listen in on a group of super-intelligent aliens whose language you don't understand, and whose spaceship only flies by Earth once an hour. It's not unlike what Harvard scientists and others are doing, except their target species, sperm whales, thankfully live here on Earth. As part of the nonprofit Project CETI (Cetacean Translation Initiative), a multi-institutional endeavor to discern the language of sperm whales, engineers in the John A. Paulson School of Engineering and Applied Sciences (SEAS) and others have led the development of a powerful listening device that adheres to whales and records high-fidelity audio and other information that's later analyzed by machine learning models. The device, called a bio-logger, collects large, high-quality datasets of whale sounds called codas, which to human ears sound like a series of rhythmic clicks, along with contextual clues like physical behavior and ocean depth. Thebio-loggeris among the first to be explicitly designed to capture data for interpretation by machine learning algorithms. Modern machine learning techniques can help uncover structured, non-human communication by identifying patterns and frequencies in the whale codas that humans can't readily perceive. The bio-logger has so far been deployed on whales off the Caribbean coast of Dominica during numerous deep-sea dives. The details of the device's design and the inspiration behind it arepublishedinPLOS One.