Dr. Fei Fei discusses the complexities of training machines to recognize objects in images. Unlike children, who can learn from a few examples, computers require millions of images to understand object concepts. She explains the need for high-quality datasets and the innovative approach of using Amazon Mechanical Turk to curate a comprehensive dataset of 15 million images across 22,000 concepts, highlighting the challenges faced in the early days of AI development.
"Our goal is we need to train machines with as much information as possible on images of objects. But objects are very very difficult to learn. A single object can have infinite possibilities that is s..."