ELI5: what is the difference between deep learning and machine learning

Machine learning is like teaching a dog a trick, while deep learning is like teaching a dog to learn new tricks on its own.

Imagine you want to teach your dog, Sparky, to sit. That's like machine learning.

• You show Sparky what "sit" means.
• You give him a treat when he sits correctly.
• You correct him when he doesn't.

Now, imagine you want Sparky to learn to fetch different objects without you telling him exactly how. That's more like deep learning.

• You give Sparky lots of pictures of different objects: balls, shoes, newspapers.
• You show him examples of "fetching" in various situations.
• Sparky looks for patterns in those pictures and examples on his own. He figures out what a "ball" looks like, even if it's a different color or size than the ones he's seen before.
• He also figures out that "fetching" involves picking something up and bringing it back.

So, both machine learning and deep learning help computers learn, but deep learning can handle more complicated tasks because it learns features and patterns automatically from large amounts of data, while machine learning usually needs more guidance. Think of it like teaching a child to identify cats: machine learning might involve pointing out features ("pointy ears," "whiskers"), while deep learning would involve showing the child thousands of cat pictures and letting them figure out the features themselves.