Scientists have succeeded in developing an artificial intelligence capable of identifying contents in an image with nearly the same precision as a human! Previously, most “computer vision analysis” programs could identify only a single object in an image. For instance, when shown a photo of a man running to catch his flight, the computer could identify “a man,” but it couldn’t recognize the man’s location or activity.
This newly developed software, a collaboration between Stanford University and Google, can discern and describe the entire scenario captured in a photo. As per the New York Times, the computer can generate “astonishingly accurate” descriptions. Such artificial intelligence could potentially be used for categorizing and organizing photos, videos, and data online.
The developed system can be trained to improve its results over time, employing a software architecture known as artificial neural networks. This system emulates the operational principles of the human brain. With this system, a computer can self-learn to process data more efficiently over time and with experience. Some of the results obtained during the initial trials are remarkably accurate, while others might evoke laughter.
Eventually, the software is intended for surveillance applications. Picture this: a security camera not merely records a crowd exiting a workplace but also identifies any attempts of theft. Like the premise of the “Person of Interest” TV show, such AI could detect and report a potential crime to authorities before it even happens. Though it sounds somewhat Orwellian, it’s seen as a significant step toward creating machines that can think like humans.
Follow us on TWITTER (X) and be instantly informed about the latest developments…
