Anil Ananthaswamy combining his background in electronics and computer engineering with his talent for making complex ideas accessible, wrote Why Machines Learn to explore how machine learning has evolved.
It covers the mathematical and computational underpinnings of AI and its societal implications.
The book begins with the foundational concept of machine learning: identifying and learning from patterns in data.
Ananthaswamy introduces the concept of the perceptron—the first artificial neural network.
The book illustrates how Bayes’s Theorem provides a systematic way to incorporate new data into existing models, laying the groundwork for probabilistic reasoning in machine learning.
It introduces Voronoi diagrams, a method for dividing space into regions based on proximity to a set of points, and connects them to the workings of the k-nearest neighbor (k-NN) algorithm.
This chapter introduces dimensionality reduction, a technique used to simplify datasets by reducing the number of features.
This chapter explores innovations that emerged well after my college years, introducing me to powerful tools like kernel methods and support vector machines (SVMs).
The chapter also explores the concept of receptive fields, the portions of an image that each neuron in a CNN is sensitive to.
The epilogue dedicates to examining the capabilities and limitations of large language models (LLMs).