SAS Institute Of Management Studies - The Power of Machine Learning: Boost Business, Accumulate Clicks, Fight Fraud, and Deny Deadbeats 

Machine learning in 20 seconds

Specialization overview: Machine Learning for Everyone

Why this course isn't "hands-on" & why it's still good for techies anyway

What you'll learn: topics covered and learning objectives

Vendor-neutral courses with complementary demos from SAS

DEMO - Exploring SAS® Visual Data Mining and Machine Learning (optional)

Deep learning: your path towards leveraging the hottest ML method

A tour of this specialization's courses

Your instructor: a rap star stuck in a nerd's body

MODULE 1 - The Impact of Machine Learning

This module covers the business value of machine learning, the very purpose that it serves. You'll see what kinds of business operations machine learning improves and how it improves them. And we'll lay the foundation: what the data needs to look like, what is learned from that data, and how the predictions generated by machine learning render all kinds of large-scale operations more effective.

Predicting the president: two common misconceptions about forecasting

The Obama example: forecasting vs. predictive analytics

The full definitions of machine learning and predictive analytics

Buzzword heyday: putting big data and data science in their place

The two stages of machine learning: modeling and scoring

Targeting marketing with response modeling

The Prediction effect: A little prediction goes a long way

Targeted customer retention with churn modeling

Why targeting ads is like the movie "Groundhog Day"

Another application: financial credit risk

Myriad opportunities: the great range of application areas

"Non-predictive" applications: detection, classification, and diagnosis

Why ML is the latest evolutionary step of the Information Age

Week 2 - MODULE 2 - Data: the New Oil - We are up to our ears in data, but how much can this raw material really tell us? And what actually makes it predictive? This module will show you what your data needs to look like before your computer can learn from it ? the particular form and format ? and you'll see the kinds of fascinating and bizarre predictive insights discovered within that data. Then we'll take the first steps in forming a predictive model, a mechanism that serves to combine such insights.

The big deal about big data

A paradigm shift for scientific discovery: its automation

Example discoveries from data

The Data Effect: Data is always predictive

Training data -- what it looks like

Predicting with one single variable

Growing a decision tree to combine variables

More on decision trees

The light bulb puzzle

Measuring predictive performance: lift

DEMO - Training a simple decision tree model (optional)

Week 3 - MODULE 3 - Predictive Models: What Gets Learned from Data - And now the main event: predictive modeling. This module will show you how software automatically generates a predictive model from data and the elegant trick that's universally applied in order to verify that the model actually works. We'll visually compare and contrast popular modeling methods and demonstrate how to draw a profit curve that estimates the bottom line that will be delivered by deploying a model. Then we'll take a hard look at both the potential and limits of machine learning ? how far advanced methods like deep learning could propel us, and yet why fundamental data requirements ultimately impose certain restrictions.

The principles of predictive modeling

How can you trust a predictive model (train/test)?

More predictive modeling principles

Visually comparing modeling methods - decision boundaries

DEMO - Training and comparing multiple models (optional)

Deploying a predictive model

The profit curve of a model

Deployment results in targeting marketing and sales

Deep learning - application areas and limitations

Labeled data: a source of great power, yet a major limitation

Talking computers -- natural language processing and text analytics

Week 4 - MODULE 4 - Industry Perspective: AI Myths and Real Ethical Risks - Machine learning is sometimes referred to as "artificial intelligence", but that ill-defined term overpromises and confuses just as much as it elicits excitement. The first portion of this module will clear up common myths about AI and show you its downside, the costs incurred by legitimizing AI as a field. Then we'll turn to the great ethical responsibilities you are taking on by entering the field of machine learning. You'll see five ways that machine learning threatens social justice and we'll dive more deeply into one: discriminatory models that base their decisions in part on a protected class like race, religion, or sexual orientation. But then we'll shift gears and balance this out by defending machine learning, demonstrating all the good it does in the world and holding its criticisms up to a higher standard.

Why machine learning isn't becoming superintelligent

Dismantling the logical fallacy that is AI

Why legitimizing AI as a field incurs great cost

Ethics overview: five ways ML threatens social justice

Blatantly discriminatory models

The trend towards discriminatory models

The argument against discriminatory models

Five myths about "evil" big data

Defending machine learning -- how it does good

Course wrap-up