Computers have already relieved their human creators of plenty of mental chores, such as doing their taxes and keeping track of their appointments. But what about reading a date’s signals at dinner?
Now, just in time for Valentine’s Day, three undergraduates at Carnegie Mellon University have applied computer technology to the science of romance with their EyeTable, an artificially intelligent dinner table that reads physical gestures and speech patterns and lets the participants know how the date is going—in real time.
Here’s how it works: EyeTable’s centerpiece is a pair of motion sensors that communicate with sensors attached to a headset worn by each participant. The table analyzes the movements and orientation of the participants’ heads—sensing whether they are making eye contact or glancing restlessly around the room, whether they’re drifting into more intimate proximity with one another or leaning apart. The headsets are also equipped with microphones that register levels of enthusiasm in the couple’s dialogue, as well as the frequency and length of awkward silences.
“We just kind of took simple elements that seemed likely to indicate the tone of the conversation and used the available technology to try to detect when those gestures were occurring,” said Dan Eisenberg, one of the machine’s creators. Mr. Eisenberg and his classmates Kevin Li and Ilya Brin invented the EyeTable for a lab course in applied computational intelligence.
But the EyeTable isn’t just an armchair analyst; it’s also a wingman. If it senses a date is going well, it might suggest an index of post-dinner activities, or tip off the waiter that the table might enjoy another bottle of wine. If it senses the date is doomed, it conveniently lists the numbers for local cab companies.
Amusing as it is in concept, Mr. Eisenberg acknowledged the practical applications of the EyeTable prototype are limited. After all, live feedback could easily have the effect of making a bad date worse—or derailing a potentially good one, since subjects behave differently when they know they are being studied. Also, headsets with motion sensors and microphones? Not sexy.
But revolutionizing the dating scene wasn’t the students’ goal. “It’s a proof-of-concept that we hope will inspire others to think about how computers can understand human emotions,” Mr. Eisenberg said. –Steve Kolowich
