AI and Machine Learning for Coders: A Programmer's Guide to Artificial Intelligence

I’m a tad suspicious about listening to books that are too deep in the weeds with code. If they’re about programming concepts, audiobooks can be suitable, but if they involve code like this one, I like to have a physical picture of the lines of code. However, I was pleasantly surprised that this book conveyed many ideas despite communicating code aurally, too. Artificial intelligence (AI) and machine learning (ML) are huge topics today. I read this book just to supplement my broader knowledge, but while reading, I found a few applications that relate directly to my work. I’m going to try out TensorFlow, the software demoed here, in my work this week. I can see where it might be a lot more effective for my users than calling remote, server-based services from OpenAI. This book features the Python-based TensorFlow as the main framework. It shows how to use it to build your own large-language model (LLM). When I started this book, I didn’t realize that it’d also demonstrate how to use TensorFlow in web situations, which is my main domain. TensorFlowJS can use an LLM built in Python to perform features in a website. This book also describes how an API can house AI/ML models from TensorFlow. Mobile devices and embedded systems like Raspberry Pi receive their own chapter, too, due to their limited processing capabilities. Honestly, I did not have super-high expectations when I read this book. I just wanted to fill in my AI knowledge from a programmer’s perspective. I did not expect to apply the knowledge directly. Instead, I found my mind engaged with new ways to think about AI – even from a non-programmer’s perspective. I’m now able to better explain how to maximize AI use to my non-IT, scientific team. And I’m really chomping at the bit to see whether TensorFlowJS can meet the needs of my latest project. It’s a total joy when a book surprises with its intellectual depth and applicability. This book did both in spades, and I’m glad I listened to it!

I’d summarize this in a sentence as “an introduction to machine learning with TensorFlow,” TensorFlow being Google’s platform for building, training and distributing AI models. I enjoyed this book a lot more than than 100-Page Book on Machine Learning, because the examples and scenarios felt more practice and grounded in the kind of interface that I would have with AI/ML in the workplace. This book was more focused on tools than theory, and went into theoretical concepts when needed to demystify the parameters/API of various TensorFlow toolsets. So far, this has been my favorite introduction to the topic, and I have ready 3-5 different AI/ML books.

I have a subscription to online O'Reilly, so I can view the book. It is pretty good as far as having practical details. It is not like it is just a collection of code with no explanation. Usually Amazon shows the table of contents, but not for this one. So here it is: Foreword Preface Who Should Read This Book Why I Wrote This Book Navigating This Book Technology You Need to Understand Online Resources Conventions Used in This Book Using Code Examples O’Reilly Online Learning How to Contact Us Acknowledgments I. Building Models 1. Introduction to TensorFlow What Is Machine Learning? Limitations of Traditional Programming From Programming to Learning What Is TensorFlow? Using TensorFlow Installing TensorFlow in Python Using TensorFlow in PyCharm Using TensorFlow in Google Colab Getting Started with Machine Learning Seeing What the Network Learned Summary 2. Introduction to Computer Vision Recognizing Clothing Items The Data: Fashion MNIST Neurons for Vision Designing the Neural Network The Complete Code Training the Neural Network Exploring the Model Output Training for Longer—Discovering Overfitting Stopping Training Summary 3. Going Beyond the Basics: Detecting Features in Images Convolutions Pooling Implementing Convolutional Neural Networks Exploring the Convolutional Network Building a CNN to Distinguish Between Horses and Humans The Horses or Humans Dataset The Keras ImageDataGenerator CNN Architecture for Horses or Humans Adding Validation to the Horses or Humans Dataset Testing Horse or Human Images Image Augmentation Transfer Learning Multiclass Classification Dropout Regularization Summary 4. Using Public Datasets with TensorFlow Datasets Getting Started with TFDS Using TFDS with Keras Models Loading Specific Versions Using Mapping Functions for Augmentation Using TensorFlow Addons Using Custom Splits Understanding TFRecord The ETL Process for Managing Data in TensorFlow Optimizing the Load Phase Parallelizing ETL to Improve Training Performance Summary 5. Introduction to Natural Language Processing Encoding Language into Numbers Getting Started with Tokenization Turning Sentences into Sequences Removing Stopwords and Cleaning Text Working with Real Data Sources Getting Text from TensorFlow Datasets Getting Text from CSV Files Getting Text from JSON Files Summary 6. Making Sentiment Programmable Using Embeddings Establishing Meaning from Words A Simple Example: Positives and Negatives Going a Little Deeper: Vectors Embeddings in TensorFlow Building a Sarcasm Detector Using Embeddings Reducing Overfitting in Language Models Using the Model to Classify a Sentence Visualizing the Embeddings Using Pretrained Embeddings from TensorFlow Hub Summary 7. Recurrent Neural Networks for Natural Language Processing The Basis of Recurrence Extending Recurrence for Language Creating a Text Classifier with RNNs Stacking LSTMs Using Pretrained Embeddings with RNNs Summary 8. Using TensorFlow to Create Text Turning Sequences into Input Sequences Creating the Model Generating Text Predicting the Next Word Compounding Predictions to Generate Text Extending the Dataset Changing the Model Architecture Improving the Data Character-Based Encoding Summary 9. Understanding Sequence and Time Series Data Common Attributes of Time Series Trend Seasonality Autocorrelation Noise Techniques for Predicting Time Series Naive Prediction to Create a Baseline Measuring Prediction Accuracy Less Naive: Using Moving Average for Prediction Improving the Moving Average Analysis Summary 10. Creating ML Models to Predict Sequences Creating a Windowed Dataset Creating a Windowed Version of the Time Series Dataset Creating and Training a DNN to Fit the Sequence Data Evaluating the Results of the DNN Exploring the Overall Prediction Tuning the Learning Rate Exploring Hyperparameter Tuning with Keras Tuner Summary 11. Using Convolutional and Recurrent Methods for Sequence Models Convolutions for Sequence Data Coding Convolutions Experimenting with the Conv1D Hyperparameters Using NASA Weather Data Reading GISS Data in Python Using RNNs for Sequence Modeling Exploring a Larger Dataset Using Other Recurrent Methods Using Dropout Using Bidirectional RNNs Summary II. Using Models 12. An Introduction to TensorFlow Lite What Is TensorFlow Lite? Walkthrough: Creating and Converting a Model to TensorFlow Lite Step 1. Save the Model Step 2. Convert and Save the Model Step 3. Load the TFLite Model and Allocate Tensors Step 4. Perform the Prediction Walkthrough: Transfer Learning an Image Classifier and Converting to TensorFlow Lite Step 1. Build and Save the Model Step 2. Convert the Model to TensorFlow Lite Step 3. Optimize the Model Summary 13. Using TensorFlow Lite in Android Apps What Is Android Studio? Creating Your First TensorFlow Lite Android App Step 1. Create a New Android Project Step 2. Edit Your Layout File Step 3. Add the TensorFlow Lite Dependencies Step 4. Add Your TensorFlow Lite Model Step 5. Write the Activity Code to Use TensorFlow Lite for Inference Moving Beyond “Hello World”—Processing Images TensorFlow Lite Sample Apps Summary 14. Using TensorFlow Lite in iOS Apps Creating Your First TensorFlow Lite App with Xcode Step 1. Create a Basic iOS App Step 2. Add TensorFlow Lite to Your Project Step 3. Create the User Interface Step 4. Add and Initialize the Model Inference Class Step 5. Perform the Inference Step 6. Add the Model to Your App Step 7. Add the UI Logic Moving Beyond “Hello World”—Processing Images TensorFlow Lite Sample Apps Summary 15. An Introduction to TensorFlow.js What Is TensorFlow.js? Installing and Using the Brackets IDE Building Your First TensorFlow.js Model Creating an Iris Classifier Summary 16. Coding Techniques for Computer Vision in TensorFlow.js JavaScript Considerations for TensorFlow Developers Building a CNN in JavaScript Using Callbacks for Visualization Training with the MNIST Dataset Running Inference on Images in TensorFlow.js Summary 17. Reusing and Converting Python Models to JavaScript Converting Python-Based Models to JavaScript Using the Converted Models Using Preconverted JavaScript Models Using the Toxicity Text Classifier Using MobileNet for Image Classification in the Browser Using PoseNet Summary 18. Transfer Learning in JavaScript Transfer Learning from MobileNet Step 1. Download MobileNet and Identify the Layers to Use Step 2. Create Your Own Model Architecture with the Outputs from MobileNet as Its Input Step 3. Gather and Format the Data Step 4. Train the Model Step 5. Run Inference with the Model Transfer Learning from TensorFlow Hub Using Models from TensorFlow.org Summary 19. Deployment with TensorFlow Serving What Is TensorFlow Serving? Installing TensorFlow Serving Installing Using Docker Installing Directly on Linux Building and Serving a Model Exploring Server Configuration Summary 20. AI Ethics, Fairness, and Privacy Fairness in Programming Fairness in Machine Learning Tools for Fairness The What-If Tool Facets Federated Learning Step 1. Identify Available Devices for Training Step 2. Identify Suitable Available Devices for Training Step 3. Deploy a Trainable Model to Your Training Set Step 4. Return the Results of the Training to the Server Step 5. Deploy the New Master Model to the Clients Secure Aggregation with Federated Learning Federated Learning with TensorFlow Federated Google’s AI Principles Summary Index

Tonight, I completed reading of Laurence Moroney’s book “AI and Machine Learning for Coders”. I enjoyed this book from cover to cover, and I am especially thankful for a few things. First, this book did a great job of explaining how tensorflow can be used to solve different types of problems, such as computer vision, natural language processing, and time series forecasting. It even went into text generation, which got my creativity going. Second, this book is written for coders, but it does explain how neural networks work at a high level. It doesn’t drop a wall of math on the reader, which I appreciated. I feel much more comfortable now, after reading this. And finally, the book ends by discussing interpretability, bias/fairness, and Google’s AI principles. I found that to be a beautiful ending. I give this five stars. The few issues I had were trivial. This book is outstanding. I’m so much better off for having read it.

Met and exceeded my expectations. Great walkthroughs and explanations on coding machine learning into different environment (using TensorFlow). Also good starter discussion on ML techniques. I especially liked the mx+b example of neural networks, very intuitive. Answered lots of questions I had about using my models, such as with JavaScript or using transfer learning. The last chapter on ethics / federated learning felt a little short (but the fact it was there was already good). I would have liked a little more on dealing with / identifying model drift and training on new data while not overwriting the weights of the old. Essentially more lifecycle stuff. Discussion of ML model deployment was all TensorFlow but like I really enjoyed it all.

Laurence is the teacher you always wish you had. He covers every aspect of TensorFlow with deep yet tangible intelligence. If you're worried about understanding the practice of machine learning and artificial intelligence, don't worry any longer. This book guides you through various aspects of the TensorFlow framework, but it also leans into the varying aspects of popular AI problems. You get the full tour in this book, from NLP coding to embedding your models in edge devices. Each section is overflowing with code samples to help you solidify what you're learning. 10 out of 10, a cornerstone of any AI bookshelf.

Laurence explains tough concept in easy way. Code in Github and website of Laurence has good examples This is my 3rd book on Machine Learning, and I have seen different approaches to explaining complex concepts like How CNN works? etc. The stye used by Laurence comes with many decades of experience to understand and I guess thats why he can explain so well. I am liking it :). I just finished Chapters 1, 2, and 3 with coding along the way. Will update again as I read and practice further. Thanks Laurence Moroney. Mahtab Syed Melbourne

If you start with this book, watch Laurence's videos on the TensorFlow/ Google Developer youtube channel and work all the examples and exercises, you will have a solid foundation in deep learning. Then you could potentially learn more from the documentation itself or by using Aurélien Géron's book. I feel other Oreilly books should be like this. Although math is almost absent in this book (hence 'for coders'), Laurence does a great job explaining convolutions and maxpooling etc.. I've seen lectures from many sources that misattribute pixel/dimension reduction to the convolution step and not the pooling layer (worse yet, lump both together) but the author clearly explains the concepts.

Fue decepcionante leerlo y ver qué solo es un tutorial de tensorFlow, realmente las explicaciones de ML y AI son muy escuetas. Si lo que quieres es aprender sobre bases de machine Learning y AI este no es el libro. Yo lo compre para preparar unas entrevistas de trabajo, realmente no me sirvió.

If you like to learn new frameworks or technologies by tinkering with practical examples, this book is excellent. It provides great coverage of data preparation and the creation, training, testing, and evaluation of models for computer vision, natural language, and time series data using Tensorflow and the Keras Sequential API.

Based on the book's description, I expected an introduction to the principles and algorithms behind AI and how to turn them into usable code. Instead, this book offers little more than a lengthy tutorial on how to use a pre-existing API called TensorFlow. It's an absolutely fine book for people who want to learn how to use TensorFlow, but a bit misleading, since as a developer I would like to know exactly what my code does, math an all, and I prefer to rely as little as possible on existing frameworks. It seems to want to appeal to a different kind of programmer.

Good, easy to follow

Non è accettabile che un libro del genere sia stampato in bianco e nero. Come da titolo, il contenuto vale 5 stelle, ma stampare un libro del genere in bianco e nero significa dare poca importanza ad immagini e grafici, quando invece in libri simili questi hanno un peso non indifferente. Stiamo pur sempre parlando di un libro dal costo di 60€.