MIT CSAIL has made a wearable device that would solve all of Sheldon Cooper’s problems. The smart wristband can tell whether a conversation is going well or you’re misinterpreting it completely #objectmagic
Engineers took Samsung’s Simband wearable and gave it an audio capture function. The device can already measure heart rate, blood pressure and skin temperature but with this ability, it will also pick up tone, pitch energy and words from a current conversation.
By taking all of them into account, algorithms can tell how a conversation is going and ultimately, the outcome you should expect. This can be especially useful for the Sheldon Coopers out there or if during an interview.
To help the system identify correctly every five second installment of discussion, the team “fed” 500 signals to the engine, letting AI choose the most important ones. Some of the findings? Long pauses are signs of hard confessions, sad stories and restless mood.
Their system can determine the success or failure of a discussion with 83% accuracy, which is pretty good considering they can make this tech work in the real world, not just in a lab.
In the near future, researchers hope to add alarms that make the device buzz differently when things are getting awkward or they are on the right track.