Listen to Today's Edition
Cats are mysterious beings. It is hard for their owners and veterinarians alike to decipher their emotions, hindering a crucial understanding of their well-being.
Now, however, help is on the way.
That’s because even though cats can fool humans, they cannot fool artificial intelligence as easily, Scientific American reported.
In a new study, a group of AI researchers and doctors developed two algorithms to establish whether a cat undergoing treatment was experiencing pain, by analyzing their facial expressions.
When being treated, cats, which are very secretive, often express pain only when it has built up so much that they cannot hide it anymore. Veterinarians, though, want to be aware of their pain early on.
So researchers, to decipher the cat’s expressions, fed the machines photographs of cats of different breeds and ages.
One of the two algorithms that was trained by the researchers to look for characteristic signs of muscle contraction, showed a 77 percent accuracy rate. It also provided an unexpected result: instead of looking at a cat’s ears or eyes, it showed it was better to focus on a cat’s mouth when trying to decipher its feelings.
The scientists concluded that AI could help note key details the human naked eye cannot.
Nonetheless, the machine only does so much as it analyzes what a cat is doing with its facial muscles, not what the cat is feeling. The researchers added that more data and human insights could improve the algorithm.
Still, researchers say the study could help improve pet care, with a knock-on effect: “When you improve the welfare of pets, you improve the welfare of people,” Stephane Bleuer, a veterinary behaviorist in Tel Aviv, who was not involved in the study, told Scientific American. “It’s like a family.”