American scientists from the University of Cincinnati have developed an artificial intelligence-based algorithm that detects human emotions during a conversation with an interlocutor by analyzing indicators such as heart rate, breathing, and sweating. The study was published in the scientific journal IEEE, writes AR, reports UNN.
Details
In their work, the authors wanted to better understand the phenomenon of physiological synchrony, in which people's respiratory rate and heart rate synchronize during communication or joint activities. This phenomenon is believed to signal sympathy for another person.
The experiment involved 36 people divided into 18 pairs. During the tests, the participants communicated with each other according to different conversation scenarios designed to evoke different emotional reactions. During the conversation, scientists recorded the biological indicators of the volunteers.
Based on the collected data, the scientists created an algorithm for classifying the type of conversation by physiological indicators. The neural network was able to identify people's emotions in different scenarios with an accuracy of 75%.
The results of the study can be applied in practice to assess the results of a wide range of communications, including the dynamics of a date or any other conversation.
Штучний інтелект знайшов величезне родовище міді у Замбії06.02.24, 02:00