A team of scientists from the Massachusetts Institute of Technology (MIT) created an algorithm that can anticipate human interactions more accurately than ever before. The artificial intelligence system, trained on 600 hours of YouTube videos and popular TV shows, can correctly decide, after just one second of a scene, whether two individuals will hug, kiss, shake hands or give a high five. It was also able to anticipate what object would likely appear in the video five seconds later.
The algorithm used systems called "neural networks," which automatically classified one of the four actions. Then those actions were merged into one, and the algorithm used it as its prediction.
When shown a video of people who are a second away from performing one of the four actions, the algorithm correctly predicted the action more than 43% of the time, which compares to existing algorithms that could only do 36% of the time, according to the study.
"Humans automatically learn to anticipate actions through experience, which is what made us interested in trying to imbue computers with the same sort of common sense," said Carl Vondrick, PhD student at MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL).
The research has implications for how machines may work with humans in the future. Computer systems that predict actions would open up new possibilities, such as robots that can better navigate human environments, predict emergency situations like falls and even headsets that feed you suggestions for what to do in different situations.
