A. P. Rodrigues

When Equations Learn

An Intuitive Guide to the Math of AI

Artificial Intelligence Isn't Magic. It's Math. And This Book Will Show You How.

Have you ever wondered what the real secret is behind artificial intelligences like ChatGPT or Gemini? How does a machine understand what we say, generate images from text, or learn to master a game?

The answer is more elegant and accessible than you imagine. It doesn't lie in impenetrable code or unreachable theories, but in a central, powerful, and surprisingly intuitive idea: that equations can be taught to learn.

In When Equations Learn, A. P. Rodrigues takes you by the hand from the first fun intuition about this concept to a deep understanding of how it works. This isn't just another book about AI; it's an invitation to understand the heart of learning.

Learn by Playing: Start with a simple and memorable "little game" that transforms a sequence of numbers and, in the process, reveals in practice the essence of what it means for a machine to "learn."

The Fundamental Building Block: Get to know the Perceptron, the simplest and most essential structure of all neural networks. You will understand its mathematical anatomy clearly and directly, without mysteries.

The Engine of Learning: Dive into the mathematics of backpropagation. See, equation by equation, how the "error" of a prediction is used to adjust the network's "weights," making it smarter with every step.

From Simplicity to Complexity: Understand how this fundamental principle scales from a single neuron to deep networks with multiple layers, and how training becomes efficient with the use of data batches.

Beyond the Black Box! The goal of A. P. Rodrigues is not just to teach a technique; it is to change your perception. It's about transforming the "black box" of AI into a crystal toolbox: transparent, beautiful, and powerful. You will not leave this read an expert in all areas of AI, but you will have something much more valuable: a deep understanding of its central pillar.