Deciphering animal language will be the next milestone in artificial intelligence

Source: World Internet Conference

For many years, artificial intelligence and machine learning have been used to analyze and translate human language, promoting the development of high-tech smart products and devices based on language technology. With the development of artificial intelligence technology, scientists have begun to try to use AI to decode the language of animals. Related research will not only help reveal the cognition, emotion and sociality of animals, but also help protect endangered species, improve human well-being and promote cross-species understanding and cooperation.

Source: pexels

Understanding animal language starts with AI+ "big guy"

When processing human language, AI may not understand the specific meaning it expresses. Even so, it does not affect the birth of AIGC tools such as ChatGPT.

So how can AI help us talk to animals? At the theoretical level, the industry believes that there are two steps: building a huge database for animal languages, and then conducting effective model design and training.

Sperm whales are the communication target of choice for scientists. This animal has a large brain, a complex social structure based on families, and a complex "language" system. Their "language" is a series of short, squeaky calls called "coda".

Source: pexels

In a paper published in 2019, scientists analyzed about 26,000 recordings to categorize coda according to the number of screams, rhythm, and more. But we're still a long way from actually deciphering the meaning of cetacean language. To that end, the project is deploying underwater robots and buoys equipped with acoustic sensors off the coast of the Dominican Republic in the Caribbean, capable of recording an estimated 400 million to 4 billion screams a year.

Using a not-for-profit model, the project brings together top cryptographers, roboticists, linguists, AI experts, technologists, and marine biologists. They agreed that only by adopting interdisciplinary research ideas and integrating professional knowledge in related fields can we have a more comprehensive, in-depth and thorough understanding of cetacean language.

What is the use of AI "translating" animal language

Professor Michael Brownstein of Imperial College London believes that deciphering animal language and allowing humans to communicate with animals will be an exciting milestone for artificial intelligence, and it will also be an in line with the development of artificial intelligence language ability. Logical next step.

Source: pexels

For example, research on marmosets can address certain diseases in humans. Marmosets are social animals, and their "vocabulary" consists of 10 to 15 calls, each with its own meaning. Studies have shown that, like human babies, young marmosets learn to communicate by hearing other marmosets talk to them, and marmosets carrying mutations associated with autism are good models for studying the improvement of the disease.

An MIT team developed an algorithm to convert frequency patterns from marmoset calls into pictures, and then passed these letter-like images to an artificial neural network to classify and recognize marmoset sounds.

"The New Yorker" magazine once wrote: "Artificial intelligence actually does not understand human language or animal language. Machines can learn a large amount of voice data in the database, connect various corpus with emotions and feelings, and use them in the human language. In the ever-changing language, we can find out the rules of expression such as syntax and grammar.”

On the road of human beings trying to establish communication with animals, the addition of AI is like a Rosetta Stone. With AI's translation of animal languages, humans can further interpret and finally decipher the secret language of animals.