ZD Tech: What if tomorrow you could translate what your cat is saying to you?


Hello everyone and welcome to ZD Tech, ZDNet’s daily editorial podcast. My name is Clarisse Treilles, and today I am explaining to you why AI can be the key to cross-species communication.

From marine mammals to small rodents, all communicate. If the question of animal language has long fascinated, modern science is working to decipher and understand the interactions of animal species.

And that’s where artificial intelligence comes in. By sifting through mass data, computers are able to find patterns that humans would take much longer to detect.

Earth Species Project

Two Silicon Valley entrepreneurs, for example, are working to decode non-human language with the “Earth Species Project”. This mission, presented as the Google Translate of cetaceans and primates, uses AI to examine the mechanisms of animal communication.

The method ? It is based on a machine learning technique, capable of learning a geometric representation of an entire language.

CETI

This is far from an isolated initiative. The media also talked a lot about the cetacean translation project, called “CETI”. This program applies machine learning and robotics to decipher sperm whale communication. To learn how they communicate with each other, researchers are trying to collect and identify their noises to match them with their behaviors.

Autonomous swimming robots and arrays of hydrophones are used on the seabed to collect audio and video data.

DeepSqueak

Finally, we can also mention the DeepSqueak project, straight out of the boxes at the University of Washington. This software, which owes its name to deep learning, is capable of analyzing the language of rodents.

Mice and rats communicate socially through a rich repertoire of ultrasonic vocalizations. Systems built from biomimetic and “deep” artificial neural networks detect and analyze ultrasound. In a nutshell, DeepSqueak was able to reduce false positives as well as analysis time, optimize automatic syllable classification, and perform automatic parsing on a given number of syllables.

And maybe one day we’ll have a Google Translate, or a Rosetta Stone for our animal friends.





Source link -97