Pictures, they say, are good ways of capturing memories.
Researchers from the Massachusetts Institute of Technology (MIT)'s Computer Science and Artificial Intelligence Laboratory, however, took this popular adage a step further by developing an artificial intelligence system which can tell just how "memorable" a photograph is.
The so-called "MemNet" algorithm can supposedly perform "at near-human levels" through "deep learning" techniques, which uses systems called "neural networks" to teach computers how to sift through massive amounts of data and autonomously establish patterns out of these pieces of information.
Principal research scientist Aude Oliva, who served as senior research investigator, explained that these techniques are similar to those employed in Facebook's photo-tagging feature, as well as Google's auto-search and by Apple's virtual assistant, Siri.
"While deep-learning has propelled much progress in object recognition and scene understanding, predicting human memory has often been viewed as a higher-level cognitive process that computer scientists will never be able to tackle. Well, we can, and we did!" Oliva said in a statement posted on the MIT website.
The research team fed the "MemNet" algorithm tens of thousands of images, each of which received a "memorability score" based on the ability of humans to remember it in experiments conducted online.
After this, the team compared the algorithm's performance with actual human subjects. The "MemNet" was found to have performed 30 percent compared to existing algorithms.
Lead author Aditya Khosla explained that this algorithm has many practical uses, including in the fields of education and advertising.
"Understanding memorability can help us make systems to capture the most important information, or, conversely, to store information that humans will most likely forget," Khosla said.
"It's like having an instant focus group that tells you how likely it is that someone will remember a visual message," he added.
The research team is also planning to develop a mobile-phone application that tweaks photos to make them more memorable.
