MIT Researchers Create a System to Measure the Pitch of Conversations

Researchers at the Computer and Artificial Intelligence Laboratory at MIT (CSAIL) and the Institute of Medical Engineering and Science (IMES) have created a system based on Artificial Intelligence that allows the detection of tone by the use of wearable devices that employ the people in those moments in which they are having conversations with others.

This system is up to more than 7% effective than other existing solutions. All the information about that system will be presented at the conference that the Association for the Advancement of Artificial Intelligence (AAAI) that will be celebrated next week in San Francisco.

The investigations have used the Simband wearable device, an existing Samsung proof-of-concept device, with which a number of variables were measured to determine the tone of the conversations. While this dressing device has been employed, researchers believe that the solution may also be applicable to other wearable devices, such as the Apple Smart Watch, in addition to others.

What has been basically sought is an objective way of measuring the tone of the conversations. To do this, the dressing device itself will collect data about movements performed, heart rate, blood pressure and skin temperature, combined with audio analysis and deep learning, so that Artificial Intelligence itself will determine the tone people are using when they are holding their conversations, pointing out whether they are sad, cheerful, or neutral.

At bottom, this solution wants to focus on those people who may not be able to detect social cues to interpret the tone of the conversations, as people with Asperger syndrome. To know more you can visit the official MIT website.


Please enter your comment!
Please enter your name here