The Songs of the Humpback Whale album was released in 1970 by Roger Payne, who was a marine Biologist. The idea of being able to decipher those eerie vocalizations seemed far-out back then.
Artificial intelligence, which has come at an extraordinary pace in the last decade, made this a realistic possibility. A group of scientists and experts embarked on a five-year effort last April to decode the language of one of the world's largest predators: the sperm whale.
The leader of Project CETI says that swordfish whales are intelligent and highly socially aware.
An Idea Worth Exploring
The project was created by a fellow at Harvard University. One day, he was listening to recordings of sperm whale codas, as the patterned clicks that these marine mammals use to communicate are called.
Intrigued, Goldwasser asked Gruber to share some recordings with her group and suggested that they set up a project to understand what those haunting, otherworldly noises actually meant. Although she meant it as a joke, she thought it was worth exploring. He asked machine learning expert Michael Bronstein if artificial intelligence could be used for that purpose.
The morselike structure of sperm whale vocalizations could be easily translated into ones and zeros, which was why Bronstein was convinced that they could be used for this sort of analysis.
The Dominica Sperm Whale Project was started by a Canadian Biologist named Shane Gero who spent over a decade studying a large clan of 40 families of sperm whales in the eastern Caribbean. He had painstakingly annotated the tracks with notes about which whales were talking, who they were with, and how they were behaving.
That was enough for a proof of concept. Using Gero's dataset, Bronstein tasked an NLP algorithm with detecting whales by their clicks and running it to a sample of recordings. The researcher's initial hunch was correct as the algorithm correctly identified the specific whales more than 94 percent of the time. A team was assembled to build off this encouraging result.
Collecting Clicks
With funding from the Audacious Project run by TED, the conference organization, and a number of other institutions around the world, CETI's team is now at work to address the project's first priority: growing the whale recording collection.
The problems for which we can train machine learning models data efficiently are those where we understand a lot about the data.
It will take roughly one billion clicks or 100 to 200 million codas to identify whale talk.
A pool of experts is developing a series of non-invasive video and audio recording devices. These include free-swimming robotic fish that can reach depths of thousands of feet and record visuals for hours; high-resolution hydrophones that can record all day long; and electronic tags that attach to the whales themselves.
Since we are studying sperm whales in their natural habitat, it is important to make sure we design the experiments in a way that does not cause disruption or long-term confusion.
Once a sufficient pool of data is gathered, the researchers will need to incorporate behavioral and social context to this basic architecture of sperm whale chatter.
There might be a sequence of clicks that these animals make prior to going hunting, or a set that occurs when a whale is sick, pregnant or tries to attract a mate.
Talking Back?
Researchers might try to chat with them if the code of whale language is cracked. This will involve developing a program that can generate vocalizations with estimated meanings, and then broadcast them to sperm whales to assess the response and whether it is predictable.
No one knows if the sperm whales will ever accept humans as a partner. He says that the whole thing is not about getting whales to understand humans.
