Engineers are trying to create computer systems with such capabilities – just as people can understand that they are looking for a dog, even if the animal is hiding behind a chair, and only the paws and tail can be seen.
Scientists have developed a computer system that can detect and identify objects in the real world based on the same visual training method used by people. According to researchers at the University of California at Los Angeles (USA), the system represents advances in technology called “computer vision,” which allows computers to read and identify optical images.
This can be an important step on the road to AI( ARTIFICIAL INTELLIGENCE )systems – computers that learn independently, are intuitive, make decisions based on logic, and interact with people more humanely. Although modern computer vision systems for artificial intelligence are more powerful, they depend on a specific task, which means that their ability to determine what they see is limited by the extent of their training and programming by people.
Even the best computer vision systems today can not create a complete picture of an object, only see certain parts of it – and can fool systems if you look at an object in an unfamiliar environment. Engineers are trying to create computer systems with such capabilities – just as people can understand that they are looking for a dog, even if the animal is hiding behind a chair, and only the paws and tail can be seen. People can also easily understand where the dog’s head and the rest of its body are, but this ability is still far from most artificial intelligence systems, researchers say.
Modern computer vision systems are not designed for self-study. They must be trained on exactly what to learn, and they usually look at thousands of images that are distinguishing the objects they are trying to identify. Computer systems can not explain their justification for determining what is in the picture. Systems based on artificial intelligence do not build an internal picture or a general model of the objects being studied, as people do, the researchers say.