ABOUT THIS BOOK
Machine Learning Logistics
How do you get a machine learning system to deliver value from big data?
Turns out that 90% of the effort required for success in machine learning is not the algorithm or the model or the learning - it's the logistics. Ted Dunning and Ellen Friedman identify what matters in machine learning logistics, what challenges arise, especially in a production setting, and they introduce an innovative solution: the rendezvous architecture. This new design for model management is based on a streaming approach in a microservices style. Rendezvous addresses the need to preserve and share raw data, to do effective model-to-model comparisons and to have new models on standby, ready for a hot hand-off when a production model needs to be replaced.
In this book, you’ll learn:
- Why successful machine learning projects involve many models
- How to use a decoy model to capture exact data inputs
- The capabilities needed in stream transport technology to support a stream-first microservice approach
- How to do accurate model evaluation, including use of the t-Digest
- Why a canary model is useful in a production setting
- How to achieve rapid and seamless deployment of new models
- The role of DataOps style of work and a global data fabric in making logistics for machine learning much easier
JOIN THE DISCUSSION
Add a comment below. What did you think about this book?, What was your favorite part?, What would you like to read next about?
Haven't read the book yet?, find it here: Machine Learning Logistics: Model Management in the Real World | MapR
See more available books in the Converge Book Club