If only Captain Ahab and the white whale could have had a heart-to-heart, things might have turned out differently.
It may be a bit late for the protagonists of Herman Melville’s classic novel “Moby Dick,” but talking to the massive creatures might one day be a reality. According to a group of scientists working on deciphering sperm whale communication, the time may come when a human-to-whale conversation will indeed be possible.
But first, they have to spend more time eavesdropping on the giants from the deep.
“The goal here is to find ways to use technology to connect us to nature by really deeply listening,” said marine biologist David Gruber RI ’18, lead scientist on Project CETI, a nonprofit that grew out of a series of meetings among scholars from various fields that began at the Radcliffe Institute for Advanced Study in 2017. The researchers, a collection of biologists, cryptographers, linguistics, computer scientists, and robotics experts, will begin by gathering the whale’s sounds and observing their patterns of behavior. In their final phase, they hope to play taped vocalizations back to the animals and record how they respond.
Gruber was virtually back on campus Tuesday to discuss the work with his colleagues and former Radcliffe Fellows Shafi Goldwasser RI ’18 and Michael Bronstein RI ’18. During an online talk moderated by Radcliffe Dean Tomiko Brown-Nagin they reviewed the inception of CETI (short for the Cetacean Translation Initiative), its progress, and what they hope the work might mean for the future of animal/ human understanding.
Gruber explained that while at Radcliffe in 2017 to design robots based on the soft properties of jellyfish, in collaboration with Robert Wood who directs the Harvard Microrobotics Laboratory, he grew interested in codas, the series of clicks sperm whales use to communicate. He was listening to a recording of the sounds in his office one day when Goldwasser, a cryptographer and Radcliffe Fellow exploring machine learning and data privacy, stopped in from across the hall. “We began discussing the sounds and she brought up the idea of using machine learning and invited me to this Radcliffe machine-learning working group that she’d organized,” said Gruber, a professor of biology at City University of New York.
More discussions followed, and in 2019 they applied for funding after studying an existing data set of whale sounds gathered by other researchers off the coast of the Caribbean Island Dominica and convening a two-day exploratory seminar back at Radcliffe. “We came together and kind of realized that this moment in time the window opened up where we think we could make significant inroads,” said Gruber.
Some of those inroads will be based on artificial intelligence, explained Bronstein, a computer scientist who was using his fellowship year to detect the spread of misinformation on social media sites with machine-learning algorithms when he sat in on the discussions about the giant marine mammals. Upon hearing Gruber and Goldwasser’s pitch about communicating with the leviathans his first thought was “It’s the craziest idea in the world.” But, Bronstein added, “The questions stuck in my head.”