Researchers from the Biology Department of the University of Copenhagen have successfully trained a model of mechanical learning to distinguish between positive and negative emotions in seven different species animal (Two types of horses, sheep, pigs, wild boars, goats, cows). This research paves the way for artificial intelligence to help us understand animal feelings.

According to the results of the study, published in the ISCIENCE magazine, the model after analyzing the acoustic motifs of animal voice expressions, achieved impressive accuracy 89.49%, signaling the first study that detects the emotional intensity of species using artificial intelligence.

By analyzing thousands of voice expressions of animals in different emotional states, researchers found basic acoustic indicators her emotional intensity. The most important factors as to whether the feeling was positive or negative included changes in duration, energy distribution, basic frequency and range shaping. It is remarkable, according to the authors, that these patterns were quite consistent in all kinds, which suggests that the fundamental voice expressions of emotion have evolving evolving.

“This innovative discovery provides strong evidence that artificial intelligence can decode feelings in multiple species based on vocal patterns. It has the ability to revolutionize the well -being and preservation of animals and in the management of livestock farming, allowing us to monitor the feelings of animals in real time, ” Indicates Elodi Brifer, an Associate Professor in the Department of Biology and one of the authors of the study.