Hands-On Artificial Intelligence for IoT

Expert machine learning and deep learning techniques for developing smarter IoT systems

Amita Kapoor

• Éditeur: Packt Publishing
• Année: 2019
• ISBN:
  9781788832762
• Édition: 1
• Sujet: Computer Science
• Sous-sujet: Artificial Intelligence (AI) & Semantics

Build smarter systems by combining artificial intelligence and the Internet of Things—two of the most talked about topics today

Key Features

• Leverage the power of Python libraries such as TensorFlow and Keras to work with real-time IoT data
• Process IoT data and predict outcomes in real time to build smart IoT models
• Cover practical case studies on industrial IoT, smart cities, and home automation

Book Description

There are many applications that use data science and analytics to gain insights from terabytes of data. These apps, however, do not address the challenge of continually discovering patterns for IoT data. In Hands-On Artificial Intelligence for IoT, we cover various aspects of artificial intelligence (AI) and its implementation to make your IoT solutions smarter.

This book starts by covering the process of gathering and preprocessing IoT data gathered from distributed sources. You will learn different AI techniques such as machine learning, deep learning, reinforcement learning, and natural language processing to build smart IoT systems. You will also leverage the power of AI to handle real-time data coming from wearable devices. As you progress through the book, techniques for building models that work with different kinds of data generated and consumed by IoT devices such as time series, images, and audio will be covered. Useful case studies on four major application areas of IoT solutions are a key focal point of this book. In the concluding chapters, you will leverage the power of widely used Python libraries, TensorFlow and Keras, to build different kinds of smart AI models.

By the end of this book, you will be able to build smart AI-powered IoT apps with confidence.

What you will learn

• Apply different AI techniques including machine learning and deep learning using TensorFlow and Keras
• Access and process data from various distributed sources
• Perform supervised and unsupervised machine learning for IoT data
• Implement distributed processing of IoT data over Apache Spark using the MLLib and H2O.ai platforms
• Forecast time-series data using deep learning methods
• Implementing AI from case studies in Personal IoT, Industrial IoT, and Smart Cities
• Gain unique insights from data obtained from wearable devices and smart devices

Who this book is for

If you are a data science professional or a machine learning developer looking to build smart systems for IoT, Hands-On Artificial Intelligence for IoT is for you. If you want to learn how popular artificial intelligence (AI) techniques can be used in the Internet of Things domain, this book will also be of benefit. A basic understanding of machine learning concepts will be required to get the best out of this book.

Genetic Algorithms for IoT

In the previous chapter, we looked at different deep learning-based algorithms; these algorithms have shown their success in the fields of recognition, detection, reconstruction, and even in the generation of vision, speech, and text data. While, at present, deep learning (DL) is on top in terms of both application and employability, it has close competition with evolutionary algorithms. The algorithms are inspired by the natural process of evolution, the world's best optimizers. Yes, even we are the result of years of genetic evolution. In this chapter, you will be introduced to the fascinating world of evolutionary algorithms and learn about a specific type of evolutionary algorithm, genetic algorithms, in more detail. In this chapter, you will learn about the following:

• What is optimization
• Different methods to solve an optimization problem
• Understand the intuition behind genetic algorithms
• The advantages of genetic algorithms
• Understand and implement the processes of cross-over, mutation, and fitness function selection
• Use a genetic algorithm to find a lost password
• Various uses of genetic algorithms in optimizing your models
• The Distributed Evolutionary Algorithms in the Python genetic algorithm library

Optimization

Optimization is not a new word; we have used it earlier with respect to both machine learning and DL algorithms, where we used the TensorFlow auto differentiator to find the optimum model weights and biases using a form of gradient descent algorithm. In this section, we will learn a little more about optimization, optimization problems, and different techniques used to perform optimization.

In its most basic terms, optimization is the process of making something better. The idea is to find the best solution, and obviously when we talk about the best solution, it means there exists more than one solution. In optimization, we try to adjust our variable parameters/processes/inputs so that we can find the minimum or maximum output. Normally, the variables constitute the inputs, we have a function called an objective function, loss function, or fitness function, and as output we expect the cost/loss or fitness. The cost or loss should be minimized, and if we define fitness, then it should be maximized. Here, we vary the inputs (variables) to achieve a desired (optimized) output.

I hope you can appreciate that calling it loss/cost or fitness is just a matter of choice, the function which calculates the cost and needs to be minimized, if we just add a negative sign to it then we expect the modified function to be maximized. As an example, minimizing 2 - x² over the interval -2 < x< 2 is the same as maximizing x² - 2 over the same interval.

Our daily lives are full of many such optimization tasks. What will be the best route to take to the office? Which project should I do first? Preparing for an interview what topics to read such that your success rate in the interview is maximized. The following diagram shows the basic relationship between Input Variables, the Function to be optimized, and the Output/Cost:

Relationship between input, the function to be optimized, and t...