[Animals] The most vocal animals in the world .. Can we learn the language of sperm whales someday?

[Love Pulse]

 

The most vocal animals in the world .. Can we learn the language of sperm whales someday?

 

Sperm whales have loud and choppy sounds that are used for social communication and to locate marine creatures. Scientists are trying to decipher this complex language.

The sperm whale is one of the largest toothed whales in the world and one of the most vocal of animals (Getty Images)

 

 

The sperm whale is one of the largest toothed whales in the world, and one of the most vocal animals on Earth. These whales make sounds known as clicks, with a pounding, intermittent, and echoing squeaks that allow them to communicate with their fellow whales, which are located a few kilometers to a few hundred kilometers away.

These whale clicks are among the loudest sounds in the world. However, it is of a high degree of complexity, which does not qualify it to be considered a complete language. Can we ever understand what the whales are saying?

 

In order to answer this question, many clues and data must be gathered to the voices of many of these clicks. This is indeed what a recent study did, which deposited its initial version on the “arXiv” website - which is concerned with publishing copies of preliminary research - on April 14th.

complex sound system
Sperm whales of the species Physeter macrocephalus have a brain six times larger than ours; This makes them complex social structures that whales use to socialize with each other. The whales may exchange these messages within 10 seconds or for more than half an hour.

According to the report published by Live Science, the members of the research team point out that "sperm whales' possession of this complex vocal system enables them - in principle - to adopt more complex vocal rules than other non-human animals."

The team has recorded many of the sounds of these whales. Then they used some machine learning systems to try to decode the sequences of those whales' vocal clicks.

Sperm whales were chosen as a model for the study; Because they emit clicks with a distinct structure similar to Morse code, which is used to transmit telegraphic information. Thus AI mechanisms may help decipher these ciphers easily.

Whales use vocal clicks as a means of socializing (Getty Images)

surface penetration
Sperm whales seem to use those vocal clicks for different purposes. As these whales dive to depths of up to 1200 meters in complete darkness, they had to develop an echo system that helps them locate marine creatures such as squid. These vocal clicks are also used as a means of social communication, but the structure of these clicks used in communication is more compact.

The team has collected and recorded nearly 100,000 sperm whale clicks over many years as part of an international initiative to decode sperm whale communications, known as the Cetacean Translation Initiative.

However, this number is never enough; In order for machine learning algorithms to build valuable inferences, they would need to collect about 4 billion audio clicks. So scientists have automated many of the channels so they can collect more sperm whales' sounds.

These channels included underwater microphones in areas frequented by sperm whales. They also used drones that can drop microphones to the surface of the water if these whales approach the surface in order to pick up their sounds.

Sperm whales have a brain 6 times larger than ours (Getty Images)

Despite the amount of data the team collected, the machine-learning algorithms still struggled to decipher it. Pratyusha Sharma, a doctoral researcher in the Computer Science and Artificial Intelligence Laboratory at the Massachusetts Institute of Technology (MIT), attributes the difficulty of analyzing voice compared to analyzing text to "the difficulty of analyzing where a word begins and ends." Thus, finding a pattern for these ambiguous and irregular words requires collecting more data.

 

These difficulties, however, make this project worthwhile, especially as it attempts to incorporate a variety of technological capabilities to be able to trace a non-human language. Which may one day help to study broader communities of organisms and knowledge of animal behavior.

Source: Live Science