Deep Learning Crash Course
Oliver Zeigermann
  • Course duration: 1h 7m
    Estimated full duration: 2h 5m
  • MEAP began November 2018
  • Publication in January 2019 (estimated)

How can you benefit from deep learning?

  • Accurately analyze customer buying habits so you can make great recommendations
  • Verify digital identity to protect customers from theft and fraud
  • Create intelligent voice assistants for speech-commanded shopping and customer service
  • Expand your customer base with automatic translation

In this liveVideo course, machine learning expert Oliver Zeigermann teaches you the basics of deep learning. This powerful data analysis technique mimics the way humans process information to identify patterns in your data and learn from them. With Oliver Zeigermann’s crystal-clear video instruction and the hands-on exercises in this video course, you’ll get started in deep learning using open-source Python-friendly tools like TensorFlow, scikit-learn, and Keras. If you’re ready to take the fast path to deep learning, Deep Learning Crash Course is for you!

Table of Contents detailed table of contents

Introduction

Why this course?

Why Machine Learning?

How good can you guess?

Basic Concepts of Deep Supervised Machine Learning

Our problem in the TensorFlow Playground

How does a neuron work?

Drawing Decision Boundaries with a single neuron

What can you do with a single Neuron?

Activation Functions

Fully Connected Feed Forward Networks

How many neurons do you need to decently separate the two classes?

How does a network learn?

Finding the sweet spot

Train a network on a complex shape

Summary

Classification with TensorFlow and the Keras API

Python Notebooks on Colab

Getting to know Notebooks and Colab

Getting to know our data

Getting familiar with Python and plotting libraries

Our first network with TensorFlow and the Keras API

What is being learned?

Evaluating our model

Understanding Generalization

Training a network with TensorFlow and the Keras API

Train the model

Making our model more general

Apply regularizations to your model

Summing up and saving our final model

Bringing your Machine Learning Model into Production

Converting the Keras model for tensorflow.js

Gluing together our JavaScript application

Make the Risk Calculator run on your machine

Alternative: Hosting your model on Google Cloud ML

Alternative: Running on a dedicated Linux server

Summary

About the subject

Deep learning is an emerging artificial intelligence (AI) technique that uses sophisticated analysis structures called neural networks to make accurate associations within a set of data. In particular, deep learning systems can learn by processing raw data without human-coded rules or domain knowledge. These systems are particularly adept at language and image classification, where a pattern may represent an abstract idea like feeling, intent, or even the general concept of what a cat or a dog looks like. These systems are also excellent for making predictions, such as how customers might behave or long-range weather forecasts. There’s also awesome potential for medical image analysis, highly-customized therapy for patients with developmental challenges, turning open surgeries into minimally-invasive ones, and better disaster recovery!

About the video

With an emphasis on simplicity, Deep Learning Crash Course teaches you to build machine learning models, the part of a system that makes classifications and predictions. You’ll also learn how to apply algorithms that train the model to improve based on the data it encounters. Your video guide Oliver Zeigermann launches your learning with a spotlight on how deep learning is different from other programming and data analysis techniques. You’ll work through a complete project and learn to use the most popular Python-based deep learning tools, including scikit-learn, Keras, and TensorFlow. All the tools are free and open source. The incredible machine learning library Keras has a minimalistic, instantly-comfortable API that handles most of the math, so you’ll get the maximum return on your time. As you work your way through this practical video course, you’ll gain skills like training a neural network, creating and executing TensorFlow code, encoding your data, and making your model more general. By the end, you’ll know how to evaluate your results, debug and improve your model, and deploy it for production.

Prerequisites

You need beginner to intermediate Python programming skills and some experience working with organized data files, such as databases or spreadsheets.

What you will learn

  • The basics of neural networks
  • Machine learning techniques using Scikit-learn, TensorFlow, and Keras
  • How to train a machine learning model and evaluate the results
  • Debugging and improving your model
  • Deployment in a production environment

About the instructor

Oliver Zeigermann is a machine learning instructor at Hamburg University of Applied Sciences. He also works as a consultant, helping to take organizations to the next level with machine learning

Manning Early Access Program (MEAP) Watch raw videos as they are added, and get the entire course, complete with transcript and exercises, when it is finished.

placing your order...

Don't refresh or navigate away from the page.