Imagine being able to decipher what birds are saying to each other or understand what African elephants are saying to each other using a smartphone. The good news is that it doesn’t seem like much time until animal speech recognition technology arrives, and in fact, it’s part of the future of Earth Species Project (ESP) technology.
ESP is a non-profit organization that Aza Ruskinthe founder of Mozilla Labs and Brit Selville It was founded by members of the founding team of Twitter. The purpose of this organization is to decode non-human communications using It is artificial intelligence.
Understanding the inner thoughts of pets can be fascinating. Of course, the benefits of understanding animal communication go far beyond listening to conversations between pet dogs and neighbors’ dogs. In fact, the ability to decipher animal communication has direct implications for the protection of the environment and our planet.
Decoding animal communication can lead to the development of tools that can aid conservation research. By knowing the undiscovered characteristics of animals, scientists will be able to discover the way certain species communicate, how they hunt, eat, communicate with other animals, and how they process information received from their surroundings.
Does a wild cat know the nature of humans? Can an elephant’s memory help pass down their life stories from one generation to another?
We can through techniques Machine learning, decoding bioacoustic data and then translating this information into natural human language, we can obtain a lot of information from animal communication. These data can contribute to human efforts to protect the environment as well as reliable scientific research on different animal species and wildlife population assessment. Of course, as attractive and innovative as it sounds, achieving this goal is really difficult.
Much of the research related to decoding animal communication is based on large language models that work in exactly the same way as the models used to improve the performance of Google Bard or ChatGPT are used. Generative AI tools are very good at understanding language; Because machine learning can understand different languages, styles and contexts well and provide appropriate answers.
ZDNet Large language models are fed enormous amounts of data during many stages of training, he writes. These models learn various inputs to understand the connections between words and their meanings. Basically, a large amount of textual data from various sources such as websites, books, researches, etc. is available for these models.
In the next step, large language models are supervised by human trainers and have conversations with them to learn different concepts better and understand the contexts of the conversation better. This step also makes large language models aware of human emotions and learn precisely how language works.
Although the language of people in different parts of the world is different from each other, but all of them make communication between people possible. Since artificial intelligence is developed based on human intelligence, it is much easier to build models to process natural human language than to build models that understand communication between animals.
The biggest challenge ESP has faced in deciphering animal communication is the lack of basic data. There is no written animal language to train these models, and different communication formats between species present another challenge.
ESP is collecting data from wild animals around the world. Researchers are also recording video and animal sounds. This data is the first step to build basic models that will provide the ability to decode the speech of a wide range of animal species.
Ruskin One of ESP’s founders says that we will probably have the technology needed to create new, productive sounds in the next 12 to 36 months, and at that time, the technology can be used to decode animal communication.
Source link
