Abstract

The aim of boosting is to convert a sequence of weak learners into a stronglearner. At their heart, these methods are fully sequential. In this paper, weinvestigate the possibility of parallelizing boosting. Our main contribution isa strong negative result, implying that significant parallelization of boostingrequires an exponential blow-up in the total computing resources needed fortraining.

 

Quick Read (beta)