A company in California wants to leverage the power of machine learning to communicate with animals.
Aza Raskin is the co-founder and president of the Earth Species Project (ESP), a non-profit organization whose goal is to decipher communication with animals through a form of artificial intelligence (AI) called artificial intelligence (AI). machine learning, which strengthens bonds with living creatures and protects them, Guardian reported on July 31.
Elodie Briefer, an animal behavior expert, co-developed an algorithm that analyzes pigs’ purrs to determine whether they are experiencing positive or negative emotions. A company called DeepSqueak assesses whether mice are stressed based on their ultrasonic calls. The CETI (Cetacean Translation Initiative) project plans to use machine learning to find ways to communicate with sperm whales.
However, ESP’s approach is very different because they do not focus on deciphering the communication patterns of one species but of all species. According to Raskin, although sign language is more likely to exist among highly social animals such as primates, whales and dolphins, ESP aims to develop tools that can be applied to the entire world. animals, from worms to whales. The impetus for the establishment of ESP came from research showing that machine learning can translate different languages without prior knowledge.
The algorithm developed by ESP expresses words through physical space. Based on multidimensional geometry, the distance and direction between points (words) describe the meaningful relationship between them (semantic relationship). For example, the word “king” is related to the word “man” in the same distance and direction as the words “queen” and “woman”.
ESP wants to create a similar expression to communicate with animals, which can be used for individual and multiple species. “We don’t know how animals experience the world, but they also have emotions like joy and sadness that seem to be similar to ours and can be communicated in communication,” says Raskin.
An Khang (Follow Guardian)