How often have you had an online conversation with someone where you couldn’t detect if the one you’re talking to was being sarcastic or serious or neither? Imagine what life is like for someone with Asperger’s or any other condition that doesn’t recognize social cues, both online and offline. Researchers from the Massachusetts Institute of Technology are developing an app for wearables that will help identify the emotions behind conversations to eventually become a sort of “coach” for those with the aforementioned conditions.
While the app is not yet ready of course for public consumption, the goal is eventually to make it available for those who would need it to improve their understanding of social cues. It was made by researchers from MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and the Institute of Medical Engineering and Sciences. They partly used the Samsung Simband which picks up high-resolution waveforms on things like heart rate and blood pressure.
The app looks at things like the vitals of the wearer, the speech patterns from the conversation and even actual audio of the conversation if there’s any. Through deep learning, the app senses the over-all “tone” of the conversation and this early, it already has an 83% accuracy, according to the researchers.
In terms of security and privacy issues, the conversation isn’t uploaded anywhere since it runs locally. However, if it’s a multi-person conversation, there might be some consent issues. Eventually, the goal is to create a consumer-level app that can be used with any smartwatch.
VIA: SlashGear
