Machine Learning Foundations for Product Managers

Duke University via Coursera

Coursera based on 548 ratings

Likes

Overview

In this first course of the AI Product Management Specialization offered by Duke University's Pratt School of Engineering, you will build a foundational understanding of what machine learning is, how it works and when and why it is applied. To successfully manage an AI team or product and work collaboratively with data scientists, software engineers, and customers you need to understand the basics of machine learning technology. This course provides a non-coding introduction to machine learning, with focus on the...

Syllabus

What is Machine Learning

In this module we will be introduced to what machine learning is and does. We will build the necessary vocabulary for working with data and models and develop an understanding of the different types of machine learning. We will conclude with a critical discussion of what machine learning can do well and cannot (or should not) do.

The Modeling Process

In this module we will discuss the key steps in the process of building machine learning models. We will learn about the sources of model complexity and how complexity impacts a model's performance. We will wrap up with a discussion of strategies for comparing different models to select the optimal model for production.

Evaluating & Interpreting Models

In this module we will learn how to define appropriate outcome and output metrics for AI projects. We will then discuss key metrics for evaluating regression and classification models and how to select one for use. We will wrap up with a discussion of common sources of error in machine learning projects and how to troubleshoot poor performance.

Linear Models

In this module we will explore the use of linear models for regression and classification. We will begin with introducing linear regression and continue with a discussion on how to make linear regression work better through regularization. We will then switch to classification and introduce the logistic regression model for both binary and multi-class classification problems.

Trees, Ensemble Models and Clustering

We will begin this model with a discussion of tree models and their value in modeling compex non-linear problems. We will then introduce the method of creating ensemble models and their benefits. We will wrap this module up by switching gears to unsupervised learning and discussing clustering and the popular K-Means clustering approach.

Deep Learning & Course Project

Our final module in this course will focus on a hot area of machine learning called deep learning, or the use of multi-layer neural networks. We will develop an understanding of the intuition and key mathematical principles behind how neural networks work. We will then discuss common applications of deep learning in computer vision and natural language processing. We will wrap up the course with our course project, where you will have an opportunity to apply the modeling process and best practices you have learned to create your own machine learning model.