Amazon SageMaker Recommender Workshop

Agenda

• {0.75} Introduction

  • {0.5} Overview of ML workflow -> decks/Recommender_Workshop_01_Introduction.pptx
  • {0.25} MovieLens data intro -> notebooks/01_exploring_data.ipynb
    • preflight: CFN template check -> cfn_setup.yaml
    • launch individual notebook instances
      • what sizes?
    • clone repo -> http://bit.ly/2wkaV0N (this repo)

• [0.5] Unsupervised ML

  • concepts
  • hands-on

• [0.25] Break

• [1] Recommendation Engine

  • [0.25] concepts
  • [0.75] hands-on
    • show movie recommender website: http://sagemaker-nab-demo.s3-website-us-west-2.amazonaws.com/
    • clone repo to individual notebook instances
    • run all at beginning of workshop
    • TBD: perform training locally?
    • during training: discussion of underlying mechanisms for SageMaker (Batch, ECR, etc)

• [0.5] Hyperparameter Optimization (PPT+Demo)

  • overview
  • demo w/ XGBoost

• [0.25] Wrap-up (PPT)

  • Other resources
  • Next steps

References

• https://grouplens.org/datasets/movielens/
• https://medium.com/@julsimon/building-a-movie-recommender-with-factorization-machines-on-amazon-sagemaker-cedbfc8c93d8
• http://sagemaker-nab-demo.s3-website-us-west-2.amazonaws.com/
• https://sagemaker.readthedocs.io/en/latest/factorization_machines.html