面向工程师的实用机器学习和AI（AppliedMachincLearningandAI

许多AI入门指南可以说都是变相的微积分书籍，但这本书基本上避开了数学。作者Jeff Prosise帮助工程师和软件开发人员建立了对AI的直观理解，以解决商业问题。需要创建一个系统来检测雨林中非法砍伐的声音、分析文本的情感或预测旋转机械的早期故障？这本实践用书将教你把AI和机器学习应用于职场工作所需的技能。
书中的示例和插图来自于Prosise在全球多家公司和研究机构教授的AI和机器学习课程。不说废话，也没有可怕的公式——纯粹就是写给工程师和软件开发人员的快速入门，并附有实际操作的例子。
本书将帮助你：
·学习什么是机器学习和深度学习及其用途
·理解流行的机器学习算法原理及其应用场景
·使用Scikit-Learn在Python中构建机器学习模型，使用Keras和TensorFlow构建神经网络
·训练回归模型以及二元和多元分类模型并给其评分
·构建面部识别模型和目标检测模型
·构建能够响应自然语言查询并将文本翻译成其他语言的语言模型
·使用认知服务将AI融入你编写的应用程序中

Foreword
Preface
Part I. Machine Learning with Scikit-Learn
1. Machine Learning
What Is Machine Learning?
Machine Learning Versus Artificial Intelligence
Supervised Versus Unsupervised Learning
Unsupervised Learning with k-Means Clustering
Applying k-Means Clustering to Customer Data
Segmenting Customers Using More Than Two Dimensions
Supervised Learning
k-Nearest Neighbors
Using k-Nearest Neighbors to Classify Flowers
Summary
2. Regression Models
Linear Regression
Decision Trees
Random Forests
Gradient-Boosting Machines
Support Vector Machines
Accuracy Measures for Regression Models
Using Regression to Predict Taxi Fares
Summary
3. Classification Models
Logistic Regression
Accuracy Measures for Classification Models
Categorical Data
Binary Classification
Classifying Passengers Who Sailed on the Titanic
Detecting Credit Card Fraud
Multiclass Classification
Building a Digit Recognition Model
Summary
4. Text Classification
Preparing Text for Classification
Sentiment Analysis
Naive Bayes
Spam Filtering
Recommender Systems
Cosine Similarity
Building a Movie Recommendation System
Summary
5. Support Vector Machines
How Support Vector Machines Work
Kernels
Kernel Tricks
Hyperparameter Tuning
Data Normalization
Pipelining
Using SVMs for Facial Recognition
Summary
6. Principal Component Analysis
Understanding Principal Component Analysis
Filtering Noise
Anonymizing Data
Visualizing High-Dimensional Data
Anomaly Detection
Using PCA to Detect Credit Card Fraud
Using PCA to Predict Bearing Failure
Multivariate Anomaly Detection
Summary
7. Operationalizing Machine Learning Models
Consuming a Python Model from a Python Client
Versioning Pickle Files
Consuming a Python Model from a C# Client
Containerizing a Machine Learning Model
Using ONNX to Bridge the Language Gap
Building ML Models in C# with ML.NET
Sentiment Analysis with ML.NET
Saving and Loading ML.NET Models
Adding Machine Learning Capabilities to Excel
Summary
Part II. Deep Learning with Keras and TensorFlow
8. Deep Learning
Understanding Neural Networks
Training Neural Networks
Summary
9. Neural Networks
Building Neural Networks with Keras and TensorFlow
Sizing a Neural Network
Using a Neural Network to Predict Taxi Fares
Binary Classification with Neural Networks
Making Predictions
Training a Neural Network to Detect Credit Card Fraud
Multiclass Classification with Neural Networks
Training a Neural Network to Recognize Faces
Dropout
Saving and Loading Models
Keras Callbacks
Summary
10. Image Classification with Convolutional Neural Networks
Understanding CNNs
Using Keras and TensorFlow to Build CNNs
Training a CNN to Recognize Arctic Wildlife
Pretrained CNNs
Using ResNet50V2 to Classify Images
Transfer Learning
Using Transfer Learning to Identify Arctic Wildlife
Data Augmentation
Image Augmentation with ImageDataGenerator
Image Augmentation with Augmentation Layers
Applying Image Augmentation to Arctic Wildlife
Global Pooling
Audio Classification with CNNs
Summary
11. Face Detection and Recognition
Face Detection
Face Detection with Viola-Jones
Using the OpenCV Implementation of Viola-Jones
Face Detection with Convolutional Neural Networks
Extracting Faces from Photos
Facial Recognition
Applying Transfer Learning to Facial Recognition
Boosting Transfer Learning with Task-Specific Weights
ArcFace
Putting It All Together: Detecting and Recognizing Faces in Photos
Handling Unknow