MIT deep-learning algorithm predicts your most memorable photos

A photograph of Taylor Swift is rated as very memorable with a score of 0.903, whereas a photograph of scenery in New Zealand’s Southern Alps is rated is rated low with a score of 0.416. Photo: MIT/CSAIL A predictive algorithm developed by scientists at MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) can predict which one of your selfies, Instagrams or Tinder profile pictures will be most memorable, or conversely, more forgettable. By identifying what parts of a photo leave the greatest impression, the MemNet brain-like computer system can predict how well a person will retain a memory of any given image. “Understanding memorability can help us make systems to capture the most important information, or, conversely, to store information that humans will most likely forget,” said CSAIL graduate student Aditya Khosla. For each image, the MemNet algorithm creates a…


Link to Full Article: MIT deep-learning algorithm predicts your most memorable photos