Hands-On Intelligent Agents with OpenAI Gym

Supervised learning

Supervised learning is similar to how we would teach a kid to recognize someone or some object by name. We provide an input and a name/class label (label for short) associated with that input, and expect the machine to learn that input-to-label mapping. This might sound simple if we just want the machine to learn the input-to-label mapping for a few objects (like in object recognition-type tasks) or persons (like in face/voice/person recognition tasks), but what if we want a machine to learn about several thousand classes where each class may have several different variations in the input? For example, if the task is to recognize a person's face from image inputs, with a thousand other input images with faces to distinguish ...