The sperm whale alphabet has been discovered, thanks to machine learning | TechCrunch

Researchers at MIT CSAIL and the CETI Project believe they have unlocked a kind of sperm whale “alphabet” with the help of machine learning technologies. The results of the study, which were published under the title, Contextual and Combinatorial Structure in Sperm Whale Vocalizations, point to key advances in our understanding of cetacean communication.

The study deals with codas, a series of clicks that fulfill different linguistic functions. What we discovered is that there is a previously undescribed variation in the structure of the coda, CSAIL director Daniela Rus told TechCrunch. We discovered that the coda types are not arbitrary, but form a newly discovered combinatorial coding system.

While whale vocalization has been a key research topic for decades, the teams behind this new research suggest they have discovered a previously unknown level of nuance among talking marine mammals. The paper notes that previous research has observed 150 different sperm whale codes.

A subset of these have been shown to encode information about caller identity and clan, he explains. However, almost everything else about the sperm whale’s communication system, including basic questions about its structure and information-carrying capacity, remains unknown.

The teams were based on the work of Roger Payne, the pioneering marine biologist who died last June. Paynes most influential work included the songs of humpback whales. It has really inspired us to want to use our more advanced technologies to want to have a deeper understanding of whales, says Rus.

The teams deployed machine learning solutions to analyze a dataset of 8,719 sperm whale pods collected by researcher Shane Gero off the coast of the small eastern Caribbean island of Dominica.

We’ll get the inputs and then adjust our machine learning to visualize better and to understand more, Rus explains. And then we would analyze the output with a biologist.

The team method marked a change from the previous analysis, which studied individual coda. A richer picture is formed when the sounds are studied in context, as exchanges between whales. Contextual details are categorized using musical terminology. This includes tempo, rhythm, ornamentation and rubato. From there, the team isolated what they refer to as the sperm whale phonetic alphabet.

This phonetic alphabet makes it possible to systematically explain the observed variability in coda structure, says Rus. We believe that this may be the first instance outside of human language where a communication provides an example of the linguistic concept of patterning duality. This refers to a set of individually meaningless elements that can be combined to form larger meaningful units, like combining syllables into words.

The meaning of these “words” takes on different meanings depending on the context. The document adds:

Our results demonstrate that sperm whale vocalizations form a complex combinatorial communication system: the seemingly arbitrary inventory of coda types can be explained by combinations of rhythm, tempo, rubato, and ornamentation features. Large combinatorial vocalization systems are extremely rare in nature; however, their use by sperm whales demonstrates that they are not exclusively human and can arise from very different physiological, ecological and social pressures.

While the breakthrough is exciting for everyone involved, there is still a lot of work to be done, first with sperm whales and then possibly expanding to other species such as humpback whales.

We decided to go to sperm whales because we had a large data set and we have the ability to collect many more data sets, says Rus. Also, since clicks form a kind of discrete communication system, it is much easier to analyze than a continuous communication system. But even Roger Paynes work showed that humpback whale songs are not random. There are segments that repeat and there is an interesting structure. We haven’t done an in-depth study.

#sperm #whale #alphabet #discovered #machine #learning #TechCrunch
Image Source :

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top