From the course: Building Recommender Systems with Machine Learning and AI

Course roadmap - Python Tutorial

From the course: Building Recommender Systems with Machine Learning and AI

Start my 1-month free trial Buy for my team

Course roadmap

“

- [Instructor] This is a pretty big course so it's worth setting the stage about how all the different parts of it fit together. To get started we'll talk about the different kinds of recommender systems, the problems they try to solve and the general architecture they tend to follow. We'll then have a quick introduction to Python for people who have some programming experience but might be new to the Python language. The next topic is evaluating recommender systems. Defining what makes a good recommendation is in itself, a complicated question and it's important to decide what you're optimizing for. We'll then build a software framework for generating and evaluating recommendations using real data and using different recommendation algorithms. Throughout the rest of the course, we'll use this framework to quickly implement and evaluate new ideas. Content based filtering will show us how we can recommend items to people just based on the attributes of the items themselves. Neighborhood based collaborative filtering gets into the classic behavior based recommendation approach pioneered by Amazon that's still in widespread use today. We then move into more modern model based methods that generally rely on matrix factorization approaches to generate recommendations. Then it's time to bust out the neural networks. If you're new to deep learning, we have a section that will get you up to speed on modern neural networks. And then we'll go through several examples of using deep learning to generate recommendations and the challenges that are specific to recommender systems when using deep learning. Next we'll scale things up to massive data sets. We'll learn how to use Amazon web services, Apache Spark, and some exciting new technology open sourced by Amazon to generate recommendations from millions of people across catalogs of millions of items. We'll then discuss real world challenges you'll encounter when deploying recommender systems for real items to real people and some of the solutions that work for them. We continue to keep it real by looking at case studies of how recommender systems work at YouTube and Netflix which ties together many of the concepts covered earlier in the course. Finally we'll look at hybrid solutions that allow us to combine many different recommender systems together into one that's even more powerful. And we'll wrap things up with tips on how you can continue your education in research into recommender systems as the field continues to evolve over time. This a very comprehensive course, covering some of the latest research and practical implementation tips in the field. And throughout the course, you'll have hands on examples all along the way to see what you've learned in action and individual exercises to allow you to practice what you've learned on your own. A quick note about how to use this course depending on your previous experience. This course is intended to be usable by anybody with some background in computer science but not necessarily people who already know Python specifically or who have prior experience with neural networks and deep learning. If you already have experience in these fields it's okay to skip these sections of the course. Nothing happens in the intro to Python or the intro to deep learning sections that later sections build upon. If you already know Python you can skip that whole section and proceed through the rest of the course just fine. And if you are already familiar with neural networks, tensor flow and recurrent neural networks, you can skip the intro to deep learning section or at least the portions of it that you're already familiar with.

Contents