Using AI to translate human language is not new, but the results that can be obtained in deciphering animal language could be stunning.
Dolphins and mice are the biggest suspects among the animals whose language we could come to understand with the use of artificial intelligence.
In a probable attempt to break the sweetest scientific news of the year, a group of scientists from the University of Washington has announced a new AI system that aims to decipher the squeak of mice.
DeepSqueak, an advanced software, can analyze rodent vocalizations and compare sounds with the behaviors observed in laboratory settings. With this process, a progressive decoding of the language of mice and other rodents can take place: the researchers hope that this technology can be useful in a vast field of medical and psychological studies.
The visual measurements thus obtained are then analyzed with the help of advanced machine learning algorithms, belonging to the same "family" as those used by current autonomous driving systems to "see" the road and the context in which the cars move without driver.
The advantage of the DeepSqueak system is to be able to "hear" sounds otherwise unheard by the human ear: "rodents have a very rich communication system that goes beyond the range of sounds that humans can hear," says Russell Marx, one of the researchers who presented the study. "Our software works to visualize all these sounds, observe their shape and structure and organize them by category."
See DeepSqueak in action
Marx and other co-creator Kevin Coffey are psychology and addiction scholars and have already made interesting discoveries in both fields: they became interested in understanding rodent satisfaction or depression signals by working with them in drug addiction experiments.
Mice, Coffey notes, are very happy at the sight of a prize (like sugar), but they show satisfaction even in particular social situations. Again: male rodents are more "repetitive" when they are between subjects of the same gender, while they make their range of expressions more complex in the presence of a specimen of the opposite sex. (Geez, like us humans! At the pub between boys only football and engines?).
We could go on and on imagining future applications of this technology, but the research team's short-term goal is to use it for in-depth addiction research.
"If scientists could better understand how substances change brain activity by inducing pleasant or unpleasant sensations," say the researchers, "we could develop much more effective addiction treatments".
Let's see what the mice say.