19 Recurrent Neural Networks
In most of this book we’ve considered every sample as an isolated entity, unrelated to any other samples. This makes sense for things like photographs. If we’re classifying an image and decide that we’re looking at a cat, it doesn’t matter if the image before or after this one is a dog, a squirrel, or an airplane. The images are independent of each other. But if an image is a frame of a movie, then it can be helpful to look at it in the context of the other images around it. For example, we can track objects that might be temporarily obscured.
When we work with multiple samples whose order matters, we call that ...
Get Deep Learning now with the O’Reilly learning platform.
O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.
