The algorithm, which translates acoustic signatures, could be used to automatically monitor the state and mood of animals, the researchers said.
Scientists believe that an AI “translator” for pigs that explains what emotions an animal feels when grunting, snorting and squealing can pave the way for better treatment on farms.
“We have trained an algorithm to decipher the grunts of pigs. Now we need someone who wants to turn the algorithm into an application that farmers can use to improve the welfare of their animals,” said Dr Elodie Brifer, who led the work at the University of Copenhagen.
Briefer and her colleagues trained a neural network to recognize whether pigs were experiencing positive (like happiness or excitement) or negative (fear or stress) emotions. To do this, scientists used audio recordings and data on the behavior of pigs in various situations – from birth to death.
In total, the analysis used the acoustic signatures of 7414 sounds from more than 400 pigs. Most of the recordings were made on farms, others came from experimental enclosures where pigs were given toys, food and unfamiliar objects to wear and explore.
The scientists used an algorithm to distinguish sounds associated with positive emotions from those associated with negative ones. They imagined emotions both in positive situations — huddling with littermates, feeding, running around, and reuniting with family — and in negative ones — fighting, castration, and waiting at the slaughterhouse.
The researchers found that there were more high-pitched squeals in negative situations. A short grunt and grunt, on the contrary, is a sign of pig contentment.
“In positive situations, the sounds are much shorter, with little fluctuation in amplitude. In particular, grunting starts high and gradually decreases in frequency, ”said Briefer.
Briefer and her colleagues believe their algorithm could pave the way for new automated animal husbandry systems that monitor sounds on farms to assess the psychological well-being of animals.