machine-learning

Machine Learning Logistics - Model Management in the Real World

  • O’Reilly book by Ted Dunning and Ellen Friedman
  • The book is free, because they are hoping it will serve as good marketing material for MapR Convergent Data Platform.
  • Focuses on the larger issue of “Machine Learning is more than just developing a model and getting it into production.”
  • Talks about the need for a holistic approach to the cross-cutting concerns of ML that we will call “machine learning logistics” and “model management.”
  • Talks about the whole system from start to finish
  • The book may be a bit dated, as it was published in 2017.
  • The book itself provides useful content. It’s not just marketing material. Every now and then, you see “tools like MapR Convergent Data Platform”, etc.

Chapter 02: What Matters in Model Management

Chapter 02

  • Logistics and model management need to be flexible to handle many different use cases and scenarios.
  • But you don’t want it to be so complicated that it is also a hindrance to people getting work done.

Ingredients of the Rendezvous Approach

  • “The rendezvous architecture takes advantage of data streams and geo-distributed stream replication to maintain a responsive and flexible way to collect and save data, including raw data, and to make data and multiple models available when and where needed.”
  • “The design strongly supports ongoing model evaluation and multi-model comparison. It’s a new approach to managing models that reduces the burden of logistics while providing exceptional levels of monitoring so that you know what’s happening.”

Chapter 01: Why Model Management?

Machine Learning Logistics Book

Chapter 1

Intro

  • “Model management” is a term for the holistic view of the entire machine learning system.
  • It’s not the cool part of Machine Learning, but it’s extremely important.
  • New practitioners are going to need some practical advice on this matter.