The wearer will know how to detect the sad emotion

MIT researchers have developed a wearable app that detects human emotions when chatting.

When you tell a friend about a new car, some will feel excited about you, while others say you are trying to show off. To help discover the true emotion when talking, a research team at the Massachusetts Institute of Technology (MIT) has built an application for wearables.

This app will collect physical data and story content for real-time analysis. Using artificial intelligence, the application can also find out the story is fun or sad and track emotional changes in about 5 seconds.

In the test, participants were asked to wear a Samsung Simband bracelet with a pre-installed MIT application. The bracelet also monitors the physical changes of the participants, such as increased temperature, heart rate or restlessness. In general, the artificial neural network can determine the emotional state with an accuracy of up to 83% (though it is unclear whether this study is comprehensive or not).

Currently, AI will evaluate the duration of the story or voice tone is a sad feeling, the rest is happy. The team hopes to soon assign more complex emotional states to the application.

"Imagine at the end of a conversation, you can rewind it and watch the moments when people around you," said Tuka Alhanai, an MIT graduate student and a member of the team. You feel the most anxiety. Our job is to develop in this direction, not far off when people live in a world where everyone has an AI in their pocket. "

This is one of the efforts from MIT's Department of Computer Science and the Artificial Intelligence Lab to study emotion detection technology . Last autumn, the team built a device to determine human emotions using wireless signals.