The Math of Neural Networks

Author Michael Benson offers the following before starting his book – ‘This book is designed as a visual introduction to the math of neural networks. It is for BEGINNERS and those who have minimal knowledge of the topic.’ For REAL beginners it is helpful to find some definitions of neural networks before beginning this intense course. For example, from the dictionary we learn ‘In information technology, a neural network is a system of hardware and/or software patterned after the operation of neurons in the human brain. Neural networks -- also called artificial neural networks -- are a variety of deep learning technologies. Commercial applications of these technologies generally focus on solving complex signal processing or pattern recognition problems. A neural network is a series of algorithms that attempts to identify underlying relationships in a set of data by using a process that mimics the way the human brain operates. Neural networks have the ability to adapt to changing input so the network produces the best possible result without the need to redesign the output criteria. The concept of neural networks is rapidly increasing in popularity in the area of developing trading systems.’Now, with a bit of very basic information such as offered above, this book is a definitive exploration and teaching guide as to how to learn neural networking, the facets of it that require concentration to ingest, and the fascination this new form of processing information addresses. Benson makes the reading as accessible as possible with ample illustrations to allow our eyes to organize the concepts into knowledge and understanding.This is not a quick read self-help book but rather a course on a subject about which many of us are uninformed. It takes time and energy to reach the end of the book, but it is very worthy time well spent.The verbiage that makes Benson’s teaching is displayed in the following definition: ‘Artificial Intelligence (AI) is an area within computer science that aims to develop machines capable of imitating intelligent human behavior. Additional Details: Neural networks are part of what is called Deep Learning, which is a branch of machine learning. The goal of Deep Learning is to move machine learning towards artificial intelligence. Machine learning is the science of getting computers to act without being explicitly programmed.General definitions section at the end of the book may just be a fine starting point for most who are absolutely new to this concept. Burt however you elect to jump into this arena the results are rewarding: it is great to gain a clue as to how communication is changing! Grady Harp, October 17

This is an excellent stand-alone book, though if you are a complete novice to the basics of what a neural network is and how it works, I would recommend first starting with Taylor’s other book, “Make Your Own Neural Network: An In-depth Visual Introduction For Beginners” which will walk you through creating your first neural network, step by step, down to explaining each line of code (using Python, but no prior knowledge is assumed).If you have at least a vague idea already of what a neural network is and how it works even if just in principle, then by all means dive right into this book which focusses, as the title suggests, on the mathematics of neural networks.However! This is not a dry maths textbook, and the explanations keep it tied closely to the topic at hand with examples aplenty, and in the same style as Taylor’s other work, very lucid step-by-step visual walkthroughs of everything, explaining each part quickly but without assuming prior mathematical knowledge (beyond perhaps the basest of concepts; nothing that should challenge anyone who has even just a high-school level understanding of maths, or perhaps even a not quite that).All in all, this is an incredibly accessible and illuminating book that I highly recommend to any who have an interest in the under-the-hood aspects of machine learning.

I got this book for my son who is currently a UTA student, who is working towards his bachelors in software engineering and is going to be writing programs for AI. I thought this would be an amazing book to purchase for him to read and incorporate his further understanding into the ideas and concepts relating to his subject of study and field of interest. I hope to get his feedback on this book, and promise that once he reads it, i will update this review with his thoughts and comments added to mine. When I told him I purchased this and wanted him to read it, he read the description and his comments were extremely positive, he said he is excited and will find time to read this sometime this week. I went through it and have to say it is above my head and I did not understand all of it, no where near it, but I did enjoy trying to learn what my son is learning about, and it gives me a small bit working knowledge so we can discuss the ideas he has as well as what he is learning in class.

If you want to read this book for its informative properties, make sure that you have the time and patience to be able to follow it. There are some complex ideas in it, and although it is presented in an easy to understand way, if you aren’t ready to learn, your eyes may glaze over fast. The good news is that there are lots of breaks in the writing and the chapters have graphs and other graphics in order to help you understand what’s going on. Another beneficial thing is that there are a lot of terms that are defined that go along with this topic, so at the very least, you’ll be able to grasp some of these vocabulary words after reading this one. Additionally, there is an extensive list of sources and articles to check out along the way, so you can see some real life examples for further review.

The Math of Neural Networks: A Visual Introduction for Beginners by Michael Taylor is a comprehensive book that details and explains neural networking. It is more than advisable to have done some research on neural networks prior to reading this book, as it is a complex subject and requires a basic understanding. This book is not to be read in one short sitting, but digested slowly, and re-read to fully comprehend the subject matter at hand. Taylor has done a good job of breaking down the information, and if you are a novice, has simplified the language and presentation somewhat to make it more accessible. Anyone who wants to learn more about Deep learning and artificial intelligence will benefit from reading this book. Highly recommend.

Really didn't do much for me, not great. Info but not put entertainingly... Done in the style of the ou numeracy books that would work...

It was really good for entry level. It should have more examples though

Compré este libro principalmente por la promesa de ejemplos prácticos con Python y TensorFlow. Ahora que lo terminé con los ejemplos puedo dar una lista de pros y contras.# Pros- El texto es altamente práctico, va directamente a las definiciones de lo que es una red neuronal *en general* y deja otros detalles culturales hasta el final.- El texto admite los lugares donde no ahonda en detalles y contiene enlaces a varios recursos adicionales en internet (artículos académicos, libros, cursos...)- Explica bien no sólo los temas a abordar, sino el orden en el que se harán, de forma que el texto es un poco más lógico que muchos textos académicos (por ejemplo, te dice cuál es la meta de cada capítulo y cómo se llegará a ella)# Contras- Pocos ejemplos prácticos. En general, el libro sólo tiene dos ejercicios prácticos al final del tema principal, uno de ellos usa Python y TensorFlow, el otro usa TensorFlow y una herramienta adicional hecha por Google.- El texto del código es significativamente más pequeño que el texto normal, tanto que tuve que subir varios tamaños de letra para poder leerlo cómodamente.- Aunque no es algo difícil, el libro asume conocimientos mínimos de Python (instalación y compilación en general)-# Pro o Contra según el lector- El texto se detiene a explicar prácticamente todas las funciones que se usan en código. Si bien esto ayuda a un principiante, para cualquiera con conocimientos intermedios de Python sólo le alentará en la lectura y es una copia de la documentación oficial de TensorFlow- El texto no ahonda mucho en detalles matemáticos. Si bien va con el tema del libro ("para principiantes") en varios lugares sólo menciona que "esto es lo más usado" sin explicar el por qué ni hacer comparaciones.- Los autores son un poco repetitivos en cuanto a los conceptos que explican. Esto fue un poco molesto para mí, pero quizá necesario para alguien más

I was confused by the amount of diverging information on neural network math in the internet, the courses and the literature (Stanford, Google, MIT, ...). Implementing a neural network was also challaging to me because of the complexity and amount of the tools available. I think the flow in the explanations of the maths from Michael gave me a simple base line to start implementation and learning from.

Simple to use examples, followed by even better simplified methods and illustrations, makes this an exciting journey in the complex field of neural networks. I highly recommend this to budding data scientists and data engineers.