5 Curves and Surfaces
In machine learning, we frequently work with various kinds of curves and surfaces. Two of the most important properties of these objects are called the derivative and the gradient. They describe the shape of a curve or surface, and thus which directions to move in order to climb uphill or slide downhill. These ideas are at the heart of how deep systems learn. Knowing about the derivative and gradient is key to understanding backpropagation (the topic of Chapter 14), and thus knowing how to build and train successful networks.
As usual, we’ll skip the equations, and instead focus on building intuition for what these ...
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.
