Machine Learning for Mere Mortals
Nicholas Chase
  • Course duration: 2h 29m
    Estimated full duration: 7h
  • MEAP began February 2018
  • Publication in June 2018 (estimated)

A great course for mere mortals!

Irfan Ullah

What can you do with machine learning?

  • Use social media data to put the right ads in front of your users
  • Predict which customers are going to leave in time to stop them
  • Real-time product pricing that reacts to competition and demand
  • Create smart SPAM filters
  • Read text and numbers from images

If you're a halfway decent Python programmer, you can learn to do all these things and much more! Machine Learning for Mere Mortals is a practical video course that takes you from clueless beginner into the incredible world of machine learning with Python and Tensorflow!

Table of Contents detailed table of contents

The basics

The basics

Machine Learning versus Artificial Intelligence

Supervised learning

Unsupervised learning

Reinforcement learning

A quick math refresher

Slope of a line

Scalars, vectors, and tensors

Matrices and matrix arithmetic

Set up your computing environment

Install Python tools

Create virtualenv environment

Install Tensorflow

The projects

Supervised learning

Supervised learning

Trend lines

Cost functions

Minimizing cost functions

Visualizing data

Using linear regression to predict values

More complicated functions

Working with matrices

Letting Tensorflow do the hard work

More Supervised Learning

More supervised learning

What are features?

What makes a good feature?

Decision trees

K-nearest neighbor

Linear classification

Making it work in Tensorflow

Creating a Spam filter

Tools and data for email classification

Classifying emails

Unsupervised learning

How clustering works

Clustering algorithms

Introducing K-means

Animating Data

Create the data and initial clusters

Assigning Points to a Centroid in K-means)

Using TensorFlow expand dims

Creating the initial K-Means clusters

Twitter trending: Download the raw data

Convert Twitter data into a corpus

Create a Bag of Words

Analyze the corpus

Introducing neural networks

What are neural networks, and how do they work

The Tensorflow Playground interface

Adding nodes to use multiple models in the TensorFlow Playground

What hidden layers are, and how to use them with TensorFlow Playground

What is the activation function in a neural network?

Using neural networks

How encoding non-numeric data works

The MNIST data set

Math doesn’t follow the real world

One hot encoding

How image recognition relates to a neural network

Making it work in Tensorflow: handwriting recognition

Softmax and cross entopy loss

Create and run the optimizer

Evaluate the results

Adding a neural network to our handwriting recognition

Adding a convolutional layer

Adding a fully connected layer

Adding a new hidden layer

Overfitting and how to prevent it

Encoding and representation

Types of data

Types of numeric representation

Text

Images

Audio streams

Analytics, stock prices, and other time-series data

Preparing data

Finding the data set

Features engineering

The mathematical way

Feature selection

Geometry of the data space

The curse of dimensionality

Ensembles

Choosing models and algorithms

Difference between an algorithm and a model

The quality of a model

Chaining together models

Making it work in Tensorflow

Downloading Twitter data

Recognizing sentiments in Twitter data

Discovering trends in Twitter data

Creating the Twitter corpus

Grouping documents into topics and determining sentiment

Optimizing performance and accuracy

Improving performance in ML routines

Using parallelization

How to parallelize an algorithm

Outliers and robustness in ML

Outliers in data

What should we do with outliers?

Detecting outliers

Robustness, overfitting, and regularization

Robustness and noise

Overfitting

Regularization

L1 regularization

L2 regularization

Other machine learning APIs

Proprietary options

Google Prediction Engine

Amazon Machine Learning tools

IBM Watson Predictive Analysis

Microsoft Deep Learning Kit

Apple AI

Open source options

Apache Spark

DeepLearning4J

Hadoop/Mahout

Google Experiments

OpenAI

About the subject

With machine learning (ML), you can predict outcomes, identify trends, and make on-point recommendations that take the guesswork out of marketing, pricing, and other key business activities. And a quick look through the job boards will tell you that machine learning has become one of the hottest job skills out there. You can't afford to miss out!

About the video

What do we mean by mere mortals? It's simple! We don't expect you to know any specialist mathematics or highbrow computer programming. If you passed college stats and you know the basics of Python programming, you're set. In this course, you'll start by learning what machine learning is, along with a quick refresher on the math you'll need, including key ML terms like scalars, vectors, and matrices. Next, you'll start working with Google's amazing TensorFlow machine learning library as you take your first steps. In your first major project, you'll build a smart spam filter. As you explore practical lessons in supervised and unsupervised machine learning, you'll learn how to fine-tune it to catch exactly what it needs to, every time.

One of the hottest ML topics is Deep Learning with neural networks. That's where this course goes next, but DON'T PANIC! You'll find examples and explanations that make this extraordinarily cool topic easy to understand. You'll build your first network, discover what makes it tick, and apply it to recognize handwriting. Along the way, you'll start to think like a machine learning developer as you learn how to choose and optimize algorithms and explore other tools you can use beyond TensorFlow.

Expert author Nick Chase brings his experience writing hundreds of articles and tutorials to the world of video, as he carefully guides you through each aspect of machine learning you need to know. He breaks down key concepts and terms so you can discuss this topic with other people in the ML biz using their own language. With this video course and Nick by your side, you'll be more than ready to develop your own machine learning applications and get real, actionable insight from your data!

Prerequisites

All you need is intermediate-level Python programming skills and basic stats knowledge.

What you will learn

  • Common machine learning algorithms
  • Working with TensorFlow
  • Using neural networks
  • Making predictions
  • Recognizing patterns in big data
  • A tour of the most used machine learning APIs

About the instructor

Nick Chase is the author or co-author of close to a dozen programming books, including Active Server Pages 3.0 From Scratch, Java and XML From Scratch, and Beginning XML. He is also an IBM Developerworks Master Author, having written more than 400 articles and tutorials on various technical topics.

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.

Excellent intro to ML and the best way to get started using TensorFlow.

Brent Faust

It is a good introduction to machine learning, especially for those without strong mathematical background.

Abayomi Otebolaku

It teaches the core principles behind the topic in a manner which is easy to follow and grasp.

Mitchell Robles Jr.