Machine Learning: A Quantitative Approach (dhl)

Updated on 4/24/2018: Examples with YOLOv3 (You only look once) - the state of the art convolutional neural network models - posted to the book's download website at www dot perfmath dot com. Instructions are also given on how to obtain YOLO's call graph and understand YOLO's implementation with the Instruments tool on macOS.

 

 =========

 

Whether you are a CS student taking a machine learning class or targeting a machine learning degree, or a scientist or an engineer entering the field of machine learning, this text helps you get up to speed with machine learning quickly and systematically. By adopting a quantitative approach, you will be able to grasp many of the machine learning core concepts, algorithms, models, methodologies, strategies and best practices within a minimal amount of time. Throughout the text, you will be provided with proper textual explanations and graphical exhibitions, augmented not only with relevant mathematics for its rigor, conciseness, and necessity but also with high quality examples for both conventional ML models and deep learning models. 

 

The text encourages you to take a hands-on approach while grasping all rigorous, necessary mathematical underpinnings behind various machine learning models. Specifically, this text helps you: 

  *Understand what problems machine learning can help solve 

  *Understand various machine learning models, with the strengths and limitations of each model 

  *Understand how various major machine learning algorithms work behind the scene so that you would be able to optimize, tune, and size various models more effectively and efficiently 

  *Understand a few state-of-the-art neural network architectures such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Autoencoders (AEs), and so on 

    

From this book, you will not only learn how machine learning works but also learn some of the most popular machine learning/deep learning frameworks such as the sklearn, Caffe and Keras/TensorFlow for doing actual machine learning work. The author's goal is that after you are done with this text, you should be able to start embarking on various serious machine learning projects immediately, either using conventional machine learning models or state-of-the-art deep neural network models.