O'Reilly logo

Stay ahead with the world's most comprehensive technology and business learning platform.

With Safari, you learn the way you learn best. Get unlimited access to videos, live online training, learning paths, books, tutorials, and more.

Start Free Trial

No credit card required

Building Recommender Systems with Machine Learning and AI

Video Description

Help people discover new products and content with deep learning, neural networks, and machine learning recommendations.

About This Video

  • Learn how to build recommender systems from one of Amazon's pioneers in the field
  • This comprehensive course takes you all the way from the early days of collaborative filtering to bleeding-edge applications of deep neural networks and modern machine learning techniques for recommending the best items to every individual user

In Detail

Frank Kane spent over nine years at Amazon, where he managed and led the development of many of Amazon's personalized product recommendation technologies. You've seen automated recommendations everywhere—on Netflix's home page, on YouTube, and on Amazon–as these machine learning algorithms learn about your unique interests, and show the best products or content for you as an individual. These technologies have become central to the largest, most prestigious tech employers out there, and by understanding how they work, you'll become very valuable to them. We cover tried-and-true recommendation algorithms based on neighborhood-based collaborative filtering and work our way up to more modern techniques including matrix factorization and even deep learning with artificial neural networks. Along the way, you'll learn from Kane's extensive industry experience and understand the real-world challenges you'll encounter when applying these algorithms at a large scale and with real-world data. The coding exercises in this course use the Python programming language. We include an intro to Python if you're new to it, but you'll need some prior programming experience in order to use this course successfully. We also include a short introduction to deep learning if you are new to the field of artificial intelligence, but you'll need to be able to understand new computer algorithms. Hope to see you in the course soon!

Table of Contents

  1. Chapter 1 : Getting Started
    1. Install Anaconda, course materials, and create movie recommendations! 00:09:06
    2. Course Roadmap 00:03:38
    3. Types of Recommenders 00:03:22
    4. Understanding You through Implicit and Explicit Ratings 00:04:26
    5. Top-N Recommender Architecture 00:05:53
    6. Review the basics of recommender systems. 00:04:46
  2. Chapter 2 : Introduction to Python
    1. The Basics of Python 00:05:04
    2. Data Structures in Python 00:05:17
    3. Functions in Python 00:02:46
    4. Booleans, loops, and a hands-on challenge 00:03:52
  3. Chapter 3 : Evaluating Recommender Systems
    1. Train/Test and Cross Validation 00:03:49
    2. Accuracy Metrics (RMSE, MAE) 00:04:06
    3. Top-N Hit Rate - Many Ways 00:04:35
    4. Coverage, Diversity, and Novelty 00:04:56
    5. Churn, Responsiveness, and A/B Tests 00:05:07
    6. Review ways to measure your recommender. 00:02:56
    7. Walkthrough of RecommenderMetrics.py 00:06:53
    8. Walkthrough of TestMetrics.py 00:05:09
    9. Measure the Performance of SVD Recommendations 00:02:25
  4. Chapter 4 : A Recommender Engine Framework
    1. Our Recommender Engine Architecture 00:07:28
    2. Recommender Engine Walkthrough, Part 1 00:03:56
    3. Recommender Engine Walkthrough, Part 2 00:03:51
    4. Review the Results of our Algorithm Evaluation. 00:03:11
  5. Chapter 5 : Content-Based Filtering
    1. Content-Based Recommendations, and the Cosine Similarity Metric 00:08:58
    2. K-Nearest-Neighbors and Content Recs 00:04:00
    3. Producing and Evaluating Content-Based Movie Recommendations 00:05:24
    4. Bleeding Edge Alert! Mise en Scene Recommendations 00:04:32
    5. Dive Deeper into Content-Based Recommendations 00:04:26
  6. Chapter 6 : Neighborhood-Based Collaborative Filtering
    1. Measuring Similarity, and Sparsity 00:04:49
    2. Similarity Metrics 00:08:32
    3. User-based Collaborative Filtering 00:07:25
    4. User-based Collaborative Filtering, Hands-On 00:04:59
    5. Item-based Collaborative Filtering 00:04:15
    6. Item-based Collaborative Filtering, Hands-On 00:02:24
    7. Tuning Collaborative Filtering Algorithms 00:03:31
    8. Evaluating Collaborative Filtering Systems Offline 00:01:29
    9. Measure the Hit Rate of Item-Based Collaborative Filtering 00:02:17
    10. KNN Recommenders 00:04:04
    11. Running User and Item-Based KNN on MovieLens 00:02:26
    12. Experiment with different KNN parameters. 00:04:26
    13. Bleeding Edge Alert! Translation-Based Recommendations 00:02:30
  7. Chapter 7 : Matrix Factorization Methods
    1. Principal Component Analysis (PCA) 00:06:32
    2. Singular Value Decomposition 00:07:06
    3. Running SVD and SVD++ on MovieLens 00:03:46
    4. Improving on SVD 00:04:34
    5. Tune the hyperparameters on SVD 00:02:02
    6. Bleeding Edge Alert! Sparse Linear Methods (SLIM) 00:03:30
  8. Chapter 8 : Introduction to Deep Learning
    1. Deep Learning Introduction 00:01:31
    2. Deep Learning Pre-Requisites 00:08:13
    3. History of Artificial Neural Networks 00:10:51
    4. [Activity] Playing with Tensorflow 00:12:02
    5. Training Neural Networks 00:05:47
    6. Tuning Neural Networks 00:03:56
    7. Introduction to Tensorflow 00:12:23
    8. [Activity] Handwriting Recognition with Tensorflow, part 1 00:17:45
    9. [Activity] Handwriting Recognition with Tensorflow, part 2 00:06:28
    10. Introduction to Keras 00:02:48
    11. [Activity] Handwriting Recognition with Keras 00:09:56
    12. Classifier Patterns with Keras 00:03:58
    13. [Exercise] Predict Political Parties of Politicians with Keras 00:09:55
    14. Intro to Convolutional Neural Networks (CNN's) 00:08:59
    15. CNN Architectures 00:02:54
    16. [Activity] Handwriting Recognition with Convolutional Neural Networks (CNNs) 00:08:38
    17. Intro to Recurrent Neural Networks (RNN's) 00:07:38
    18. Training Recurrent Neural Networks 00:03:21
    19. [Activity] Sentiment Analysis of Movie Reviews using RNN's and Keras 00:11:02
  9. Chapter 9 : Deep Learning for Recommender Systems
    1. Intro to Deep Learning for Recommenders 00:02:19
    2. Restricted Boltzmann Machines (RBM's) 00:08:03
    3. [Activity] Recommendations with RBM's, part 1 00:12:33
    4. [Activity] Recommendations with RBM's, part 2 00:07:11
    5. [Activity] Evaluating the RBM Recommender 00:03:44
    6. [Exercise] Tuning Restricted Boltzmann Machines 00:01:43
    7. Exercise Results: Tuning a RBM Recommender 00:01:15
    8. Auto-Encoders for Recommendations: Deep Learning for Recs 00:04:27
    9. [Activity] Recommendations with Deep Neural Networks 00:07:08
    10. Clickstream Recommendations with RNN's 00:07:23
    11. [Exercise] Get GRU4Rec Working on your Desktop 00:02:43
    12. Exercise Results: GRU4Rec in Action 00:07:52
    13. Bleeding Edge Alert! Deep Factorization Machines 00:05:49
    14. More Emerging Tech to Watch 00:05:15
  10. Chapter 10 : Scaling it up
    1. [Activity] Introduction and Installation of Apache Spark 00:05:49
    2. Apache Spark Architecture 00:05:13
    3. [Activity] Movie Recommendations with Spark, Matrix Factorization, and ALS 00:06:03
    4. [Activity] Recommendations from 20 million ratings with Spark 00:04:58
    5. Amazon DSSTNE 00:04:41
    6. DSSTNE in Action 00:09:37
    7. Scaling Up DSSTNE 00:02:15
    8. AWS SageMaker and Factorization Machines 00:04:24
    9. SageMaker in Action: Factorization Machines on one million ratings, in the cloud 00:07:39
  11. Chapter 11 : Real-World Challenges of Recommender Systems
    1. The Cold Start Problem (and solutions) 00:06:12
    2. [Exercise] Implement Random Exploration 00:00:54
    3. Exercise Solution: Random Exploration 00:02:18
    4. Stoplists 00:04:48
    5. [Exercise] Implement a Stoplist 00:00:33
    6. Exercise Solution: Implement a Stoplist 00:02:23
    7. Filter Bubbles, Trust, and Outliers 00:05:39
    8. [Exercise] Identify and Eliminate Outlier Users 00:00:45
    9. Exercise Solution: Outlier Removal 00:04:00
    10. Fraud, the Perils of Clickstream, and International Concerns 00:04:34
    11. Temporal Effects, and Value-Aware Recommendations 00:03:31
  12. Chapter 12 : Case Studies
    1. Case Study: YouTube, Part 1 00:03:43
    2. Case Study: YouTube, Part 2 00:07:04
    3. Case Study: Netflix, Part 1 00:03:59
    4. Case Study: Netflix, Part 2 00:03:56
  13. Chapter 13 : Hybrid Approaches
    1. Hybrid Recommenders and Exercise 00:02:54
    2. Exercise Solution: Hybrid Recommenders 00:04:18
  14. Chapter 14 : Wrapping Up
    1. More to Explore 00:02:31