An Iterative Map-Reduce-Update runtime executes a series of iterations, each of which first calls map with the
required arguments, then performs a global reduce aggregation, and lastly makes a call to update.
Example: Convex Optimization A large class of machine learning—including Support Vector Machines,
Linear and Logistic Regression and structured prediction tasks such as machine translation—can be cast as
convex optimization problems, which in turn can be solved efficiently using an Iterative Map-Reduce-Update
approach 14 »
required arguments, then performs a global reduce aggregation, and lastly makes a call to update.
Example: Convex Optimization A large class of machine learning—including Support Vector Machines,
Linear and Logistic Regression and structured prediction tasks such as machine translation—can be cast as
convex optimization problems, which in turn can be solved efficiently using an Iterative Map-Reduce-Update
approach 14 »
